It’s almost 2020… and for someone like me, born in the 1970s, that still feels like a weird sentence to say. A date that always felt closer to science fiction than everyday life.
But, here we are, in the future, a place that had promised me flying cars and robots to complete every household task imaginable. It’s safe to say that on that particular front technology has fallen short, but one glimmer of the future still remains.
Voice First, Voice Assistants and Voice Controlled Smart Homes are one technology promised by the likes of Tomorrow World that does appear to have a short term future in the real world. Even if its growth has been slower than some may have imagined.
Back at the tail end of 2017, I made 5 predictions prevailing to the future of Voice and Smart Speakers and now, two years on, I thought it would be interesting to check the progress of those predictions and see if my optimism about the future of voice was misplaced.
Let’s take a look…
2017 Prediction One: “Your Mum Will Buy a Smart Speaker”
In other words… Smart Speakers would make the leap up from the Early Adopters stage of the “Innovation Life Cycle” and into the Early Majority phase.
Back at the tail end of 2017 is was expected that 55% of the US population would own a Smart Speaker during 2018. That figure hit 25% early in 2019 which, despite failing to meet those lofty predictions, still makes the adoption rate of Voice Technology faster than any other technology in history.
Smart Speakers are no longer just for the tech-savvy few who like to be first in the queue at the Apple Store and more and more people are using “Voice” in their day to day interactions with technology.
My prediction that voice interaction would become “The new normal” in 2018, however, seems to still be a way off. More people are using more Smart Speakers than ever before but for many, they are still a novelty item rather than our preferred source of information and entertainment.
If I was to revise my original prediction two years on I now think that Smart Speaker ownership may actually fall away. More and more technology is being integrated with Voice Control and Voice Assistants (more on that later) meaning there is less of a demand for those little white tubes in the corner of the room.
Use of voice will continue to rise but its presence in our homes will become less overt which will, in turn, will mean using it will feel like a more natural part of everyday life to everyone… even your mum.
2017 Prediction Two: Smart Speakers Will “Talk Smarter”
It’s generally accepted that making communication with Voice First devices as frictionless and as human as possible will assist adoption and, although there is progression here, there is still much to do.
Developers are experimenting further with the voice used within their voice experience to make it feel more natural. Replacing the Alexa voice simulation with a more natural-sounding VO which does improve the human “feel” of the interaction but this is not a practical option in many scenarios, not is it a solution to some of the bigger challenged.
Users are still reporting frustrating and clunky interaction with Smart Speaker’s largely down to the complexity of interpreting the myriad of unguided potential commands, variety of regional accents and plefora of phrases that can be used to ask for just a single function. This continues to be a huge challenge for developers.
Interestingly, research shows that voice assistant’s accuracy in answering voice queries has fallen back in the last 12 months: Research from Perficient reported that Amazon Alexa’s accuracy fell from 90% to 79% and other platforms experiencing similar drop-off. Whilst this could appear like a step backwards it’s important to look at the potential root causes of the drop-off. It’s very likely that accuracy decreases as we demand more from these devices. AI and understanding IS improving which means that when Alexa and Co can’t quite understand what’s being asked instead of returning a negative result they are taking a guess! This is backed up by the stats which show a rise in the percentage of questions the devices attempt to answer.
This one is a work in progress. Voice platforms are not only constantly looking at live feedback from customers in terms of how their Voice Assistants handle queries, but there are fundamental changes taking place in terms of how the software behind such platforms operates and souces information.
On top of these developments, Amazon has continued its crowd-sourcing model to build Alexa’s knowledge base by launching “Alexa Answers” allowing customers to fill in Alexa’s knowledge gaps themselves – like a voice Wikipedia.
There have certainly been improvements made but there is still a long way to go before communication with Voice First platforms feels “human”.
2017 Prediction Three: “More High-Quality Content Will Emerge”
Two years on and I’m still on the hunt for that killer Alexa Skill! Content is one area in which voice has struggled.
There is yet to be a true breakthrough in functionality for anything beyond audio delivery and voice searches. Whilst these are both great applications of voice it feels, to me at least, that the technology has so much more to offer.
There are plenty of developers who are consistently pushing the boundaries of what Alexa and other platforms can do but I am yet to find that one function, game or application that voice can deliver better than anyone else.
Patience is a virtue here. Voice tech is still a baby (think of it as the world wide web in the early 90s) and the more experimentation and exploration that takes place the closer we will get to voice finding its true purpose.
We are starting to now see audio content that is specifically designed for Smart Speakers rather than just delivered by the medium. Both Amazon and Google are slowly replacing news content that would have previously been delivered via synthetic voice with better-produced audio (and video) from trusted audio partners.
I expect this trend to continue with both more content creators entering the market and brands creating high-quality audio content as part of their marketing strategy.
The way the technology operates and interacts will be fundamental to the adoption of voice but it will be the great content that keeps users coming back day after day and we are yet to find that hook.
2017 Prediction Four: “Discovery Will Change”
Here we have seen change. We had to see change!
There was something fundamentally backward about logging onto a screen-based web portal to find the Voice App you needed to complete a task (ala the Alexa Skill Store) and Amazon were quick to address the issue.
There have been two key changes here:
First the ability to “Enable” a skill with an invocation term (ie. “Alexa, Open Domino’s Pizza” opens Domino’s Pizza Skill whether or not it has been physically enabled or not) and secondly the arrival of “Suggested Skills” where Alexa will suggest a third party Skill to complete a requested function (ie. “Alexa book me a taxi” would offer you the Uber Skill as an option).
Both these additional functions have had positive effects on discoverability for Alexa Skills but both still have barriers. To enable a skill with voice command you must already know the name of that Skill, which then relies on the developer marketing that product. Whilst the ability to recommend Skills is only really useful for functional requests rather than applications such as games or content.
In reality, these challenges are no different from those already adopted by the internet as a whole; similar challenges faced that have been solved with the rise of search engines. In my view, it is inevitable that we will also see similar solutions with voice search and voice SEO becoming the main player in terms of discovery over the coming months and years.
Just as we do now with Smart Phones and Desktop browsing, we will find our desired content with direct requests to our Voice Assistants which will then return the most suitable results. With Voice SEO also in its infancy, it will be interesting to see how content creators, brands and news outlets position their content so it is top of the pile when it comes to these searches. With voice, being first return is vital and those who master Voice SEO first could be in for a big win.
2017 Prediction Five: “Voice Assistants Will Be EVERYWHERE.”
Voice Assistants may not be everywhere yet… but they are coming.
Earlier this year Amazon announced a whole new raft of devices from Microwaves to Fashion Rings to Earbuds that would contain Alexa technology and take ‘her’ out of Smart Speakers and into the wider world.
This is still very much a work in progress. The Amazon tactic seems very much to throw as much tech at the wall and see what sticks and whilst some of their Alexa-enable hardware will find a market (I’m quite keen on the earbuds), others will fall by the wayside.
This Spring’s ‘Smart Audio Report‘ from NPR and Edison research showed that whilst still popular, Smart Speaker growth was starting to slow in 2018/19 with 70% of owners saying they were not considering purchasing a second device. For me, this is partly down to Voice Assistant integration into other products. Whilst many products currently use voice as a point of difference, as more people become familiar with the technology, it will soon become a commercial necessity appearing in everything from Kettles to Headphones to Entertainment Systems which is bound to have a knock-on effect on the sales of stand-alone devices.
What I find interesting is whether the voice platform we use will continue to be dominated by the hardware in which is it installed? Will, for example, an Apple device be limited to using Siri or will we be given the free choice to install and use our preferred assistant. More likely the value of a preferred assistant will be leveraged to tie us into a particular brand with the convenience of having one interconnected voice assistant making us more likely to buy into a particular technology ecosystem.
Voice “On the Go” for me is very much the sweet spot for Voice Tech. I’ve found myself using Voice more and more when driving when the convenience of hands-free becomes essential and there are no social barriers around using voice which I think hints at future use.
We all carry this technology with us, everyday day in our pockets, in the form of Smart Phones. So, the technology is already EVERYWHERE. It now needs to take the step up from gimmick to daily habit and this will continue to grow like the other elements, such as “Content”, “Discoverability” and “Communication” develop.
I think it’s pretty safe to conclude that I was way off with my 2017 predictions. Although it’s also fair to say that there has been development in all the areas mentioned above I was overly optimistic as to the pace of growth.
I’ve noticed increasing frustration with Voice First technology recently, and maybe even a little lost faith, as it doesn’t move forward with the pace that perhaps some had hoped. This is more to do with the perceived potential than the actual speed of development.
Stats like the much-quoted (and false) “50% of search will be voice by 2020” have done nothing to quell these expectations and its easy to forget that this is a technology very much in its infancy.
Not only is the tech still developing at a rapid pace but the way in which users interact with Voice is changing rapidly and my ONLY prediction for 2020 is this will continue.