In the beginning there was...
Don’t worry, I’m not going to talk to you about Genesis, but I do want to take you back in time:
In the beginning there was conversation.
Before print marketing, before radio ads, before TV ads, before search marketing. I guess in the advertising world we would call it “word of mouth”.
Voice is the oldest modality in the marketing world.
And that is how I want you to think of voice search for the rest of this article. As a modality, not as a channel, nor app.
So old in fact, that voice technology is not a new concept.
In 1769, Austro-Hungarian author and inventor Wolfgang von Kempelen created a speaking machine.
Known as the “speech synthesizer”. In its third and final form, this machine consisted of many artificial parts such as kitchen bellows, the reed from a bagpipe, plus a fake mouth and nose. Pretty impressive for the time period!
In the end, Kempelen’s speaking machine was able to converse complete phrases in French, Italian, English and German (German requiring more skill from the operator. I think we can guess why).
Cool, so what does this have to do with modern day voice search?
Well Kempelen’s research led him to some fascinating discoveries, some that we are now seeing reemerge with modern day digital assistants. These were:
- “Listeners often ignored mispronunciation altogether (a phenomenon later explored by researchers in the field of cognitive science)”. A field heavily centred around artificial intelligence and linguistics. Sound familiar?
- “Kempelen believed that people were more forgiving of the errors made by his machine due to the frequency of the reed and vocal tract resonant length he chose to use, which create a resonance much more like a young child, than that of an adult.” Source
Both of those points are interesting, as when it comes to voice technology, a big hurdle is the ability of a machine to understand context, tone, sarcasm etc…
The second point in particular probably has more weighting than you might first think.
Voice search is a conversation, so there is bound to be preference in how the assistant sounds. This will likely be subjective person to person, but a study by Voicebot.ai found the following:
- Users prefer a human male voice
- Users prefer a female synthetic voice (Cortana from Halo comes to my mind here)
- Both men & women prefer human voices over synthetic voices
Read into that what you will. There will no doubt be further studies done on preference this decade.
What it does tell us though is that voice preference will go a long way into establishing trust with a user.
Akin to people in the 1700s being more forgiving of a child’s voice, we will and are seeing that people in the 2000s may interact differently with digital assistants based purely on how they sound.
I felt it was important to preface this post with a bit of background information on voice technology and conversational semantics, as at the end of the day, voice search is based around human conversation.
These assistant devices are not linear chat bots, they are designed around natural language and turn based interactions, much like a normal face to face conversation.
But therein lies the challenge, can digital assistants deliver a human experience that is also fully optimised for search?
Where is voice search at (as of 2021)
Welcome to the 2020’s, where your shopping will be done in AR and invoked by voice.
It won’t be long until we see this conversion path: Voice Search > AR Product > Sale
I really do believe that.
But who cares what I think. What are the actual figures?
“50% of searches will be voice searches by 2020”. I’m sure you’ve heard this one before.
First cited by comScore in 2018; it is a tired and now incorrect statement that is part of every blog post about voice search (so I had to include it).
It was a prediction. Not a far fetched one. But I do see that figure thrown around as a stat. It’s not.
What does resonate more from the comScore article I feel, is their introduction:
“The future of search is voice and personal digital assistants“
I’ve highlighted “personal digital assistants” as that is the second way I want you to think about voice search.
These smart devices are assistants with voice capability. It’s literally in the name of one of the biggest voice search devices on the planet, the Google Assistant.
They are designed to be assistive. To compliment your existing tech stack & to be used in every day moments.
This approach may change the way you think about voice search. For me it definitely shifted my mindset, while also placing more emphasis on core aspects most are concerned about when it comes to voice:
- How do I optimise for discoverability (voice SEO)?
- Will it change SEO strategies?
- Why would people use voice instead of traditional search?
- Is voice search growing?
- How does voice change user behaviour?
- How does voice work hand in hand with the web?
These questions are easier to answer when you think of voice in the two ways I described:
- A modality. An interaction type triggered by a voice action to surface search results
- An assistant to complete tasks. To cut time. To make your life easier. To provide services. To enhance user experience.
Usage & ownership statistics
Use cases, usage frequency and ownership are our primary concern when optimising for voice, or building a digital assistant environment for a brand, as knowing these is a great starting point for planning a voice search marketing strategy.
50% of searches…maybe one day that figure will reign true.
A proxy for a metric like that is ownership. Ownership of digital assistants also gives us insight into their actual usage frequency.
Ownership & sales statistics:
- 84% of Americans who are 18 or older, own a smart speaker. That is roughly 60 million people. That is a 3% increase from 2019 (+7 million owners).
- 82.9% of Americans own a smartphone. 56.4% of those use a voice assistant on their phone. 34.7% own a smart speaker – voicebot.ai
- Since 2018, smart speaker ownership in the US has grown by 11.8%
- 55% of US households are expected to own a digital assistant by 2022
- Voice is expected to be a $40 Billion dollar channel by 2022
So we know that ownership is on the rise. Tick. Those are big numbers.
- A study done during this pandemic has shown that daily usage of voice commands have risen by 5%, up to 25% using voice several times per day. Couple that with 27% using voice once per day, that means for the first time ever, more than half of smart device owners are using voice commands at least once a day.
- 27% of the global online population is using voice search on mobile – Google
- 51% of those who currently use voice search to shop, use it to research products – Navar
- 39.8% of smart speaker owners use voice to stream music. “Hey Google, open Spotify”
- 29.4% of smart speaker owners ask their assistants questions daily.
- 7.3% of those who own a digital assistant use voice to search for product information daily. 4.9% then use voice to make purchases daily.
- 25% of consumers have already made at least one purchase via voice
Now we also know that daily use is on the rise. Tick.
What we see here though is that ownership does not match frequency.
Some people use Google voice search or Alexa voice search everyday, while some people probably bought a Google Home or Amazon Alexa as a novelty, asked for a few jokes and haven’t touched it since.
But there are emerging interaction types that are related to marketing, SEO strategies and branding, as seen with commerce coming onto assistant platforms.
This is a hurdle that voice search faces. The ownership is there. The frequency is growing. But the projections of usage are not hitting forecasted targets.
Use cases. The biggest hurdle voice search faces right now is the current limit on interaction types.
What to take note of is that voice search is part of the new mobile experience. Further interaction types will come while the usage of voice for online commerce will continue to rise.
This will lead to a change in the way people search for content, discover content and digest content.
So with that said, how does one best approach voice search as an SEO strategy?
Changing how people search
Voice changes how people search.
It changes both where they search from and the format of information they receive.
As voice searches are more conversational in nature, one of the best voice search optimisations you can do is optimising for informational queries.
This means a change in the content you serve on your site, as that content needs to match the user intent of the query.
SEMrush published a study in 2019 which found that “70%” of all answers returned from voice searches occupy a SERP feature (60% of those being a featured snippet).
Optimising for these traditional SEO features will help ensure you get better rankings for voice searches and increase your organic reach.
Quick rundown of featured snippets for SEO:
Featured Snippets, often referred to as position zero, are rich results where the information is pulled from third-party sources, such as content on a website.
Google displays this information on top of all organic results.
This is done to provide users with quick and concise answers.
Voice search fits into this perfectly as informational queries often generate featured snippets.
Who, what, why, when, how etc…
Google will then read the featured snippet results out loud back to users on the Assistant.
The use of natural language via voice means that informational queries are more inline with how users are interacting with assistants.
- Searches are longer
- Searches are more conversational
- Searches are less like “computer language”.
The lack of prepositions in traditional desktop searches are far less prevalent in voice search:
- Desktop: “Sydney Hotels”
- Voice: “What are the best hotels in Sydney to stay at”
So much so that Google states 70% of searches on the Google Assistant use natural language – Source
This has a big impact on how we conduct keyword research and the resulting on-page optimisations, as voice search means we need to answer informational intent in order to have any chance of ranking in the featured snippet.
So answer the informational queries! It’s the best way to earn featured snippet rankings. Great for SEO, great for voice search.
A greater emphasis on local search
Local SEO results also tie in nicely with how people search on digital assistants.
In particular, the explosion in “near me” terms are also a core component of optimising for voice search.
Here are a few pointers:
- Mobile voice search results are 3x more likely to be local pack results than featured snippet text. This is because consumers are three times more likely to search locally when searching by voice. – Source
- Get smart with keywords that may be relevant to someone searching for things that return local results. “Hotel near Hyde Park”. “Pubs in walking distance from (my hotel in Sydney)”. etc…
- Pay close attention to quick actions linked to your GMB listing. Reviews, website, call, book now & get a quote buttons are all part of the data set the Google Assistant will pull from. Ensure they’re correct! Ensure you are also tracking this data.
- Be ready for last minute lookers. This goes back to informational queries. But the fastest way to get information quickly is via a voice search (I can order a pizza in 8 seconds on my a Google Assistant). Are there quick localised questions you can answer on your website? “How do I get to x from my hotel”. “Show me Italian restaurants in Darling Harbour”. etc…
- Note that when a user searches for a local query or something “near me”, 83% of those customers who visit shops in store have performed that local search within a week.
You may have all these bases covered already from your everyday SEO efforts.
But remember: Your Alexa or Google Home is simply an entry point that uses voice activation. Much like Google is an entry point that uses text activation.
Here’s a great example of that. Check out these Google Assistant results:
Local map pack on the Google Assistant
GMB result on the Google Assistant
All this information is pulling from your GMB listing. A point that reinforces how voice search is just another modality.
What else do you notice from these images though? See those tabs at the bottom?
This is another sneaky way to get hyper local keyword and content ideas that are already deemed relevant for voice search results. Here’s what the rest of those suggestions look like for the “Shangri-la Hotel, Sydney”:
“Spa”, “Altitude”, “Blu Bar”, “valet parking cost” – Some nice information here. Importantly, these are keywords that do not appear as related results in a desktop search.
“People also ask” and “searches related to” show different results. So it could be easy to miss these!
Use the assistant to find some more keyword ideas!
So, where is this all headed?
The simple answer to that question is: The seamless integration into everyday life.
You see this already with many common items:
- Basically every smart phone has voice activation
- Your TV remote has voice activation
- Your car, your Xbox, your fridge, your lightbulbs…
The foundations for this technology are everywhere and becoming increasingly integrated into everything.
How long until every car on the road has an inbuilt digital assistant? Not long at all, I say. The future of voice search is “on the go”.
This article highlights the future of voice assistants in greater detail: https://medium.com/voiceui/the-future-of-voice-on-the-go-416112fb15db
The integration into cars alone will dramatically change how users search for and consume information.
Behavioural flows will change and give rise to a search experience that picks up where you left off, no matter what device or activation modality you are using.
What I began to realise is that digital assistants, particularly the Google Assistant, is the glue that holds all other Google products together. It sits on top of their tech stack. You can access absolutely everything via the Google Assistant.
As such, it makes sense that Google intends for people to use Google voice search to access all touch points of their accounts.
So picture this:
“Hey Google, book me a flight to New York on August 4 for 2 weeks”
As it stands right now in 2021, that search wouldn’t automatically complete that action for you. There’s too many steps involved, right?
But in the near future, the whole process of booking a flight from Sydney to New York, including transport to and from the airport, your hotel in NY, and even a dinner reservation during your stay, could be automated in one search.
In fact, Google has recently started testing “Memory” for the Google Assistant.
I think you can guess what this means… but in short, Google can and will be able to store your personalised information in the assistant itself, across everything.
Meaning that if your preferred transport method to get to the airport is Uber, Google will know. If you fly with Qantas, Google will know. If you stay with Hilton Hotels, Google will know.
This is all on top of Google already having access to the entire Google tech stack (contacts, email, calendar etc)
So now when I say: “Hey Google, book my a flight to New York on August 4 for 2 weeks“, my result will return the entire journey:
- The Google Assistant books my flight with my preferred airline on August 4 and the return flight 2 weeks later
- The Google Assistant books my Uber to and from the airport
- The Google Assistant books my preferred Hotel in New York based on previous history
- The Google Assistant sees in my calendar that I have a catch up with a friend booked on August 7, and books a table at a restaurant
- It could go on…
Voice search is a fascinating topic. It’s a field of technology that is not solely related to marketing, but one where I think it will change the marketing landscape entirely this decade.
If you enjoyed this post, I would greatly appreciate a share. Also, if you have any questions at all, on any part of this post, let me know by leaving a comment!
Check out our other posts
Plenty of more digital marketing tips, tricks & guides here!