The power of speech

Grappling with the voice gatekeepers

WPP's Atticus Awards celebrate the best and most original thinking from people across our network.

This year’s Atticus Journal contains 23 thought-provoking essays, covering everything from the changing consumer environment and how to boost brand value to the rise of chatbots and the impact of AI on our purchase decisions and voting habits.

This Grand Prix-winning article looks at the ways in which voice is set to transform people’s interaction with technology, and what brands can do to thrive in a voice-activated world.

Advertisers will need to reappraise relationships with voice assistant providers as voice user interfaces (VUIs) strengthen their role as consumer gatekeepers.

It seems unlikely that the interruptive model of paid advertising will translate easily to voice, as users demand a more streamlined experience. Google Home customers in the United States kicked up a fuss when Google Assistant gave out unsolicited information about a new Disney filmⁱ. Google quickly removed the ad.

With the future of traditional advertising in doubt, brands can reach consumers over a VUI by building services for Echo, Home or DingDong. While these services offer companies the opportunity to create a branded experience over the platform, they hold limited promise for direct consumer engagement in the future.

There has been limited uptake by users thus far. According to a 2017 report from VoiceLabs, there are more than 7,000 skills on the Alexa platform but only 31% of these have more than one customer review, suggesting the majority of applications aren’t being downloadedⁱⁱ. Users dislike the particular commands required to make the skills function properly. What’s more, the technology is currently not able to deliver a unique voice that’s also interactive, giving brands little scope to differentiate themselves from the voice assistant (as well as from their competitors).

$2 billion of sales were forecast to be driven by digital assistants in 2016 aloneⁱⁱⁱ.

Some skills stand out, such as the Uber skill, which allows Alexa users in the United States, United Kingdom and Germany to order and cancel rides or to check a request status. Uber has a loyal following and those using voice want to order from Uber particularly, rather than any taxi service the voice assistant might choose.

But many situations don’t require a branded solution, nor do users think in these terms. Instead, they have a problem, such as a grease stain on their clothes, and they want the voice assistant to tell them how to remove it before going back out to their dinner party. No one has the time or inclination to follow the steps to activate the Tide stain-removal skill specifically, for example. Any dependable advice will do.

Joseph Evans, senior research analyst at Enders Analysis, predicts, “You’ll have a layer of skills by developers and they will present themselves to the Alexa central level, saying, “I can do X, I can do Y, I can do Z.” And then when you have a user request, Alexa will do some kind of machine learning and say, “Okay, what they said means that they want Y,” right?” The user will not interact with the brand directly.

With these services likely to recede into the background of the overall voice assistant experience, and with the interruptive ad model being questioned, brands face a future in which a few key VUI providers act as gatekeepers to consumers, who rely on their voice assistant to deliver solutions to their problems and to purchase products on their behalf.

As the market for voice technology matures, brands may have to pursue a number of strategies to ensure that voice assistants surface their brand as that single result. Currently, there’s still much debate about how this may play out.

Recommendation and algorithm optimisation

A key challenge in a world intermediated by voice assistants will be ensuring your brand or content is chosen by the assistant. Algorithm optimisation will become the new SEO.

While we have seen many smart speakers with integrated visual interfaces, voice technology works best when there’s only one answer, according to 80% of global regular voice technology users. As the market for voice technology matures, brands may have to pursue a number of strategies to ensure that voice assistants surface their brand as that single result. Currently, there’s still much debate about how this may play out.

Here we consider three possibilities: paid recommendations, the affiliate model and algorithm optimisation. In some circumstances it may be possible for a voice assistant to offer a “paid recommendation” to a user. For instance, asking Alexa to buy more washing powder could lead to the (paid) suggestion of an alternative offer. The key to this model’s success will be user trust that the suggestion is right for them and not just lucrative for the voice assistant provider.

An alternative might be an affiliate model, similar to price comparison services today. The voice assistant could deliver a single recommendation in response to a query and then take a commission on a subsequent lead or sale. As of now, this approach does appear to be palatable, with 57% of regular voice users saying, “I don’t mind if [a local voice assistant] takes a commission from a purchase made by voice as long as the deal is good for me.”

In this scenario, brands that are positively reviewed by journalists and other social influencers will improve their chances of being recommended by the VA. A third option gaining traction is “assistant optimisation,” also known as “algorithm optimisation.” Much as with search engine optimisation (SEO), businesses will be able to affect the likelihood of the voice assistant recommending their brands through the structure of their digital content assets. In a voice-enabled world, ensuring your content is chosen by the assistant will become ever more critical.

Join the Internet of Things

Brands should consider adding voice interactivity as and when these capabilities expand to new devices.

Voice technology is moving rapidly beyond mobile phones and smart speakers. Alexa has already been built into dozens of smart devices. Social robots around the world such as Olly, Jibo, RoBoHoN and Musio are an alternative to smart speakers. Voice-activated products such as the kids’ digital storybook The Snow Fox or the controversial My Friend Cayla doll offer a more interactive and enriching experience than their basic equivalents. Meanwhile, the number of commercial settings that integrate voice-activated services is growing. In hospitality, a team of robots at the Henn-na hotel in Japan checks in guests and offers concierge services, while the JW Marriott San Antonio in the United States is piloting the Echo Dot in its hotel rooms. The Dot will provide guests with information about restaurants, room service and directions.

Brands need to think about how they can make their own physical assets voice- enabled, whether through the integration of a voice assistant like Alexa, or their own voice technology. For retailers, this could include embedding voice assistants in display units or changing rooms. We found that 52% of global smartphone users already want a voice assistant that can communicate with the store, for example to help them to navigate and to find products more easily.

Connected packaging also presents an opportunity. Two thirds (67%) of global smartphone users agree with the statement “I like the idea of being able to ask my products questions about their provenance.” At CES 2017, we saw the launch of talking packaging in the form of Cambridge Consultants’ AudioPack concept for drugs and medical devices. Media themselves are increasingly likely to become voice-enabled. Using voice, consumers will be able to engage as and when they want with services such as Instreamatic. AI, which is an interactive audio platform that allows advertisers to create voice-enabled audio ads. Audio is also being made searchable. Israeli company Audioburst records radio shows and transcribes soundbites called “bursts.” This content can then be tagged with keywords and discovered online and via Alexa.

As the TV becomes voice-enabled (Alexa is now built into the latest Amazon Fire stick; Google Home allows users to play video on their TVs through Chromecast using voice command), viewers may also use their voice assistants to interact with TV advertisers. Sky has also recently integrated voice recognition into its remote. Deutsche Telekom and Orange are collaborating on a smart assistant called Djingo, which will allow people to use voice to control their Orange TVs throughout Europe, the Middle East and Africa when it launches in 2018.

As brands turn to voice, they will start to amass an audio dataset that will present opportunities to glean considerable insight via voice analytics. What can you learn about your customers from their tone of voice and how they talk to you? Can you apply sentiment analysis to the voice recordings to understand how customers feel? What’s more, brands that build their own voice-enabled products and content can position themselves outside the voice gatekeeper ecosystem and take one step closer to winning back control of consumer engagement.

As brands turn to voice, they will start to amass an audio dataset that will present opportunities to glean considerable insight via voice analytics. What can you learn about your customers from their tone of voice and how they talk to you?

Are you being served?

Brands that develop useful content can meaningfully engage with their consumers via a new channel. 

There are two routes a brand can take when developing a one-to-one experience. One is to build a chatbot for a messaging platform such as Facebook Messenger, WeChat, Line or Kik. Brands should be thinking about the circumstances in which chatbots can be voice-enabled. It is possible to envisage successful text-based chatbots delivering an even more flexible experience for customers by adding voice capabilities. Already Microsoft’s Xiaoice, which has more than 40 million users in China and Japan, has launched as a voice-enabled chatbot on WeChat^iv.

The second route to consider is creating a skill, Conversation Action, or DingDong service. When contemplating this investment, brands must first decide what the software will accomplish. The characteristics of some of the early examples of successful skills for Alexa are rooted in their simplicity and ability to offer real value in a natural, straightforward way. A user who activates the BMW skill can ask Alexa about a scheduled trip, find out what time to leave, and send the destination to the vehicle. Amsterdam-based Triggi’s Alexa skill allows users to control Nest, Netatmo, Philips Hue and other smart devices, intending to “make smart things smarter.”

Services that reduce friction in people’s lives are likely to appeal to the 52% of regular voice users globally who cite convenience as one of the main reasons for using voice technology. Ultimately, voice platforms must understand the needs of the user and deliver on these through a process that is simple and intuitive.

Find your voice

Brands can strengthen consumer relationships by defining their (literal) voices.

Our neuroscience research gives an early indication that speaking to a brand delivers a deeper emotional connection than interacting with it through type or touch. When people asked a question involving a brand name, their brain activity showed a significantly stronger emotional response compared to people typing that same brand question.

The act of saying a brand name appears to strengthen the pre-existing emotional associations to a greater degree than typing it. This emphasises the need for brands to craft the sound of their own voices.

Consumers are already looking for greater variety in voices. Just under three quarters (74%) of regular voice technology users globally believe brands should have unique voices and personalities for their apps or skills, and not just use the assistant on smartphones.

Some of the early skill pioneers have started to use different voices to reflect the service they offer. The BBC News skill, for example, uses a pre-recorded presenter’s voice. Eighth Note, a voice-activated version of the smartphone game Flappy Bird, which originated in China, takes a different approach. The game character’s movement is controlled by the volume of the user’s voice, illustrating how the brand has considered the player’s voice rather than its own, to achieve engagement. Eighth Note has gone viral in Asia, where Chinese YouTube personality Jing Jing has accumulated 6.5 million views on a video of herself playing the game.

In time, we will see more personality, beyond the sound and tone of the voice, once AI capabilities allow responses to be truly interactive and adaptable. Our research has shown that people have strong preconceived notions about what a brand should sound like – globally 62% of smartphone users say their voices and personalities should be unique. Hit the right notes and you can be music to your customers’ ears. 

Short-term brand implications

What marketers can do to prepare for a voice-activated world in the short term:

1. Consider how voice could genuinely augment the touchpoints on your consumer journey. How could a voice interaction add value to or remove friction from the consumer experience?
2. Learn the rules of engagement in conversational commerce. Build a chatbot and deepen your exposure to the types of conversations consumers want to have.
3. Identify your consumers’ Deciding Factors. What are the cultural and technological sticking points in your market that must be handled carefully in order to launch your voice proposition successfully?
4. Experiment with voice-user interfaces. Test and learn, for example through an Alexa skill or a DingDong trial. Explore how you can provide utility to your customers or drive new behaviours.
5. Review whether your search activity is optimised for voice. Does your keyword strategy capitalise on the long tail of conversational search terms?

Long-term brand implications

What marketers can do to prepare for a voice-activated world in the medium to long term:

1. Enhance your privacy credentials. Review and develop your personal data privacy policies to reflect local market concerns, as attitudes to privacy in relation to voice vary considerably by market. Position yourself as the brand to trust wherever you do business.
2. Use radio or interactive audio ads to develop your brand’s voice. What do people want to hear when they speak directly to your brand?
3. See what you can learn from how customers talk to your brand. Try using voice-analytics software to detect how your customers really feel.
4. Re-evaluate your PR efforts. Getting your brands recommended by respected journalists and other thought leaders could be your best bet for staying relevant in the era of voice and the affiliate model.
5. Forge strategic partnerships for voice integration. Work with retailers and service businesses to distribute your voice-activated products and services into their environments.

 

---

^{i. Chris Welch, “Google Home is Playing Audio Ads for Beauty and the Beast,” The Verge, March 16 2017, bit.ly/2rL0iAy
ii. VoiceLabs, “The 2017 Voice Report,” January 15 2017, bit.ly/2jJlyk7
iii. Finbarr Toesland, “Voice Search and Chatbots Are Transforming Commerce,” Raconteur, September 8 2016, bit.ly/2mIu4iT
iv. Jerry Salandra, “China, WeChat and the Origins of Chatbots,” Chatbots Magazine, March 12 2017, bit.ly/2qTEQK9}

 

Download WPP's Journal of Original Thinking, Atticus Volume 24 PDF 4.5MB