Time to start experimenting with voice assistants

2018 has been a big year for voice assistants and smart speakers. According to a recent report by Deloitte, 12% of UK consumers now own a speaker, like the Amazon Echo, with an in-built digital voice assistant, and 81% have used a voice assistant on a smartphone, laptop, TV or other device. The technology brings with it a plethora of new ways for brands to engage their audiences, with the possibility of big rewards for those pioneers who can become a go-to destination for voice assistant users in their market.
Look who’s talking
While voice assistants have been embedded in our smartphones for years (Siri was released as an app for iOS in February 2010) and smart speakers with digital voice assistants have been on the market since late 2014, consumers have thus far been slow to adopt the nascent technology. While the promise of instantaneous, almost effortless interactions with a wide range of services and smart appliances is appealing an age characterised by instant gratification, a lack of understanding about what to do with voice assistants, and suspicion of the tech giants that operate them, have been keeping consumers at bay.
As usual with new technologies, younger consumers are adopting voice assistants at a faster rate, but they also tend to use their digital helpers less frequently than older adopters, According to a recent study by PwC, it’s because they fear looking silly in public.
But in spite of these concerns, adoption of voice assistants is gaining pace, with the number of installed smart speakers predicted to reach 100 million by the end of 2018, and YouGov reporting earlier this year that smart speaker ownership in the UK doubled in just six months. Regardless of whether the tipping point arrives this year or next, it’s clear that voice assistants present an important new terrain on which brands can engage consumers.
How people are using voice assistants
The same YouGov poll found that the most frequent actions performed by users on their smart speakers were playing music (71%), answering general questions (58%), setting alarms or reminders (49%), getting live news and weather updates (47%), and interacting with other smart devices (34%).
Much is made of the potential for home automation in the marketing of the two most popular smart speakers, Amazon’s Echo (which runs the voice assistant Alexa) and Google Home (which runs Google Assistant), so it’s unsurprising that people are using them to control lighting, central heating and media playback through associated apps (Skills for Alexa, Actions for Google Assistant).
But the most popular actions also highlight the limitations of voice assistants. Google puts a positive spin on it:
When people talk to their Google Assistant, they’re usually trying to get something done” says Scott Huffman, VP Engineering, Google Assistant. “Assistant queries are 40 times more likely to be action-oriented than Search, with people asking for things like ‘send a text message,’ ‘turn off the lights,’ or ‘turn on airplane mode.’
Why do we think this is happening? For many tasks, particularly while you’re on the go, it can be much easier to get things done through voice. I can say “turn on the lights and play some music,” without having to worry about which app I need to open. Even for basic things like creating a calendar invite, I don’t have to look down at my phone or interrupt what I’m doing, I can just say “create an appointment for noon on Saturday.” These seem like small things, and they are. But they illustrate what makes voice so unique—the technology allows me to complete a task in a way that feels natural. The more we can build these types of experiences, the closer we get to an ideal Assistant.
Another way of interpreting this is that voice assistants just aren’t very good for more complex, multi-step activities, like browsing and comparing product options or different sources of information on the web.
How brands are reacting to the rise of voice
So far, the most successful incursions by brands into the world of voice assistants centre around simple, oft-repeated activities. Home automation, media streaming, journey planning and news/sports apps top the lists of must-have Alexa Skills, along with apps from services like Ocado and Uber. Adding items to your weekly shop and calling a taxi are easy to accomplish with simple voice commands, whereas shopping apps like Uniqlo IQ which provides “a variety of experiences by which to find clothes, including hourly product rankings, searching by occasion type, finding items featured in monthly magazines, and even discovering products based on today’s lucky color, based on your daily horoscope” seem more like a PR exercise than a genuinely useful tool. (We’re in the middle of a barrage of publicity stunts involving voice assistants – see, for example, the Burger King ‘OK Google’ ad and The Human Test by Channel 4.)
Whatever the experience you’re looking to deliver to consumers via voice assistants, taking into account the limitations of the technology and the context of the user is critical to success.
The memory challenge
When using a graphical interface, a user’s possible actions are clearly laid out in front of them. With voice assistants, a user has to remember what actions are possible and the commands they need to use to initiate them. An interviewee in a recent Slate article recounted their experiences with the Amazon Echo:
The voice-driven operating system can also make using Alexa’s skills (essentially, its apps) a challenge. “I’m a bit nostalgic and I wanted to listen to The Shadow, the old 1930s radio show,” Swenson remembered. “Lo and behold, they had a skill for that. The thing is, though, I feel like I need a notecard by the [Echo] to give it those commands. Because you do have to use somewhat specific language with it to open up a particular skill. I couldn’t just tell it, ‘Play Shadowradio podcast.’ ” Ferro’s experience with a skill designed to keep her household on top of whether the cat had been fed went about the same way (poorly). “It was very specific about what you had to say to get it to work, which ended up being a problem for my roommates,” she said.
Some developers, like the creators of Life Bot, have experimented with working around this by creating a cross-device onboarding process:
“When you first set up the skill, Life Bot asks for your phone number. The app then texts a welcome message to your phone and asks for your name. After you reply, it will return with suggestions of things to try. Having your hand held like this through the onboarding process makes it a lot easier to get started with Life Bot, compared with other voice apps. Plus, now that Life Bot has your phone number, the app can connect with you when you’re away from Alexa, too.”
The creepiness challenge
Voice assistants, with their always-on microphones and ability to access and process a user’s most sensitive information, justifiably raise privacy and security concerns with many consumers. Security researchers have pointed out exploits involving frequencies inaudible to the human ear, which can be used by attackers to mute and then carry out all kinds of operations on smart speakers, and reports of Echo and Home devices unexpectedly sending user’s sensitive info to random contacts are common.
All of which means brands, especially when asking users to surrender personal data, have to exercise caution, as Zoe Lawrence of Kantar points out in a recent Forbes interview:
“…as with any new technology, brands need to enter into the environment carefully. It’s such a personal space that any brand presence needs to feel completely appropriate and non-invasive. We saw that 66% of Singaporeans are concerned that companies will listen into their conversations. So anybody developing a service for voice needs to appreciate and address these concerns.”
The context challenge
While voice assistants can be used on the go, via iOS and Android apps, we’ve already encountered evidence that people are wary of using voice commands in public spaces. Figures from surveys of US consumers suggest that smart speakers tend to be placed in the living room (45.9%), kitchen (41.4%), bedroom (36.8%) and home office (10.9%). Designing a successful voice assistant experience is therefore largely about solving problems that your audience will encounter in these contexts (as well as in the car, where looking silly or being overheard is less of a concern).
Experiment for success
Finding the sweet spot for voice assistant apps – providing something genuinely novel or useful, which can be triggered by easily remembered commands, which is appropriate to the user’s context and which doesn’t come off as creepy or insecure – is no mean feat. Taking a lean, experimental approach and informing development with real user feedback as early as possible is therefore essential.
Manifesto recently helped Cancer Research UK investigate whether Amazon Echo could be used to disseminate cancer information to one of their key audiences – cancer patients and their close friends and family – in an engaging, appropriate way. The joint team rapidly prototyped three different proof-of-concept propositions during the course of a single day, to both demonstrate to senior stakeholders that the route held promise, and to test with real users before moving into development.
For the charity, this paves the way to an additional route for users of their information support services to access content, ensuring that they cater to the needs of users who want to engage in ways that suit them, that are intuitive, and that don’t compromise their security or privacy.
Brands like CRUK, making early forays into the world of voice assistants, can expect several misfires before hitting upon the right formula but, if the current rate of adoption continues, can also look forward to a significant source of audience acquisition and engagement in an increasingly fragmented and personalised media landscape.
While voice assistants aren’t likely to render web browsers and mobile apps redundant, they are going to quickly usurp those channels for many consumers in certain contexts. If those consumers and contexts are relevant to your brand, then you need a strategy for voice.
Leave a reply