Audio branding is nothing new. From the familiar three tones of NBC-TV to the whisking-away sound of emailing in Microsoft Outlook, we’re used to associating sounds with brands. Since virtual assistants and voice technology have come to the forefront, brands need to realize that this first encounter is the most important one of the interaction. When you say “Alexa,” you know you’re talking to Amazon, and you know that “Hey Siri” awakens Apple’s voice-enabled software. But how can users get to your world?

Your voice roadmap is a critical extension of your brand. Some experts advise brands to partner with a known platform, like adding Alexa or Google to automobiles or appliances. But for brands that want to retain control over their brand and user experience, creating a custom wake word is the way to go.

Branded wake words enable brands to deepen user engagement, control the user experience and own their customer data.

For Motorola, the catchy “Hello Moto” phrase wakes its voice user interface on many of its smartphones. Mercedes-Benz owners are already summoning their in-car voice technology to do things like set a radio station, temperature, and seating position by saying, “Hey Mercedes...” And of course, a breezy “Hey, Pandora” sparks a personalized musical experience that’s a huge differentiator from its competitors. Connect the customized wake word with your brand itself, rather than inserting another brand wake word into the mix, and increase product name affinity.

You’ve probably had experience creating user personas — the composite traits of a typical customer of your brand. You give them photographs, names, ages, and a back story. Have you ever tried going the other way and created a persona that represents your brand? If your brand were a person, who would it be? Would it be a serious and professional colleague, a friendly, chatty helper, or a funny, clever friend? What’s its name? What gender is your brand? What does the tone and pitch of its voice sound like? Shoe companies often hire sports stars to be the voice and face of their products because customers relate to characters most like themselves, or like their ideal selves.

Austrian chatbot provider Onlim advises its customers to understand their consumer base and design the voice assistant to fit the customer personae. Consider the voice gender, age, race, and personality. The smartest direction to take would be to craft your voice assistant based on a diverse set of customer data and explore options where users can select their preferred voice. This way, you can avoid potential biases and design an assistant that resonates with your audience and feels natural to them.

“For a lot of brands, especially big brands, you already have a brand experience your customers recognize. So it's really extending that same experience to another venue, through voice. It should always start off with, what's the business goal and who are you today and is this going to be multichannel or via a new channel,” advises Heidi Culbertson, technologist, VUX designer and the founder & CEO of Marvee.

Pay attention to vocal characteristics when choosing a voice user interface to ensure that the character is right for your brand. Once you’ve done some market research to determine who your audience is, test some synthesized voice personas until you find one that matches both your audience profile and the image you want to project.

Today, synthetic voice creation requires audio recordings by professional voice actors in partnership with linguistic experts. But if you’re looking to consistently create recordings of large volumes of personalized text, speech synthesis is a great choice and a quicker way to do this. Brands can also use a company like the Acapela Group to create a voice with only a few minutes of high-quality audio recordings and the associated text transcription of the audio samples (text-to-speech).

Text-to-speech voice synthesis demo on the Acapela website

“Conversations are social interactions,” said Lauren Golembiewski, CEO and conversational and voice interface designer at Voxable. “Users often personify conversational interfaces and project human qualities onto them.” People refer to both Siri and Alexa as “she,” and talk to the interface conversationally. That quirk of humans—the same instinct that makes us see faces in random shapes—is key to the success of a voice assistant.

“We don’t want to talk to a computer. We want to talk to someone who understands us, is empathetic toward us, and who talks like we do. To ensure the VUI has intuitive conversational interactions with users, designers must establish the VUI’s voice, tone, and persona,” Golembiewski continues. “Voice is the type and quality of the words the VUI uses. Tone is how those words modulate. And Persona is the character that embodies and informs the voice and tone. (These) are the brand foundations that drive what language designers and writers use in prompts, responses, and information the VUI generates.”

For Ananya Sharan, product manager of Pandora’s Voice Mode, it’s always been about helping people find what they want, more easily. “From the beginning, keeping our listeners in mind, we wanted the voice experience to be effortless. You want it to be natural, conversational. Our vision was that the voice assistant is like that music expert friend that really knows you.”

“We wanted people to be able to ask for things like, “Play me something awesome,” adds Sharan.

Your voice user interface needs to be the literal voice of your brand. It should speak with confidence and authority, keep responses brief, and provide the right amount of information. It should be friendly, familiar, and knowledgeable, and it should be able to expect what we might ask next. Dailey of RAIN suggests building the “perfect persona” by storyboarding the values of your brand and how those tenets could be represented vocally. “Can the audience relate to (the voice) or does the VUI sound like it’s the audience’s child or parent (and if so, is that a good or bad thing?) Selecting a persona is just as important as selecting a brand ambassador and should go through just as stringent a vetting process.”

A challenge just as prominent as that of selecting a persona, is appropriately communicating and maintaining brand voice and tone. Daily adds, “Similar to how you inject your brand voice into messaging, marketing, and other communications, you will need to appropriately script your desired voice, and corresponding tonal adjustments, into your dialogue flow - and on top of that, will need to make it feel authentic and natural.”

An added layer of complication is that while many marketing channels are static in their messaging, voice is not, and requires you to account for what can be hundreds or thousands of interactions all maintaining a consistent voice.

Since we already know that people aren’t always comfortable with conversing with machines, it’s essential that the voice sounds natural and ordinary. The conversation should flow naturally. Forcing the user to press the button every time they talk can be cumbersome and unnatural.

Source: Google

Golembiewski, an expert in both conversational design and product design, often speaks on this topic with other creators of voice user interfaces. “The VUI’s discovery, usability, and user satisfaction increase when it has a consistent voice that modulates based on the situation and users’ emotional states along with a well-defined persona,” she advises. “Designers can build upon the established voice, tone, and persona by refining word-usage rules and writing style guides. It is important to establish writing examples for the VUI when documenting brand guidelines.”

Dailey agrees. “If your scripted dialogue isn’t something you’d say out loud, consider adjusting,” she said. “You can test this by “reading lines” with a partner. Dailey also recommends further anthropomorphizing your brand voice with words and phrases like “thank you,” “got it,” and “please” to personalize conversations and speak to the user in a way they’re used to from talking with other humans.

“Scripts are a key component of building a voice experience, as they quite literally guide the experience from point A to point B. Just like in any other kind of script (movies, TV, plays, etc), character consistency and development is important to ensuring fluidity and can increase the potential for a deeper connection with your audience.” adds Dailey. “To maintain a consistent voice throughout, try building a character profile to define limitations of voice and then account for areas of tonal fluctuations throughout the script. Outline, via a storyboard, the values of your brand and how you think those could be communicated or represented, the sense or mood you want to put out as well as get back from your users, and more. Additionally, storyboard out a variety of interactions you expect to have with your users to determine the specific tonal adjustments you’ll want to make against each nuanced situation.”

Dailey advises that in order to imbue some of the same characteristics that we as human beings maintain when conversing, it is critically important to adapt a variety of anthropomorphic qualities in your voice experience, such as:

Natural, authentic language:
Write scripts the way you’d speak naturally. If your scripted dialogue isn’t something you’d say out loud, but rather something you’d read, consider adjusting. You can test this by “reading lines” with a partner.

Contextual relevance:
Pick conversations back up right where they left off to show the user that you’ve not forgotten the progress they’ve made. This creates a level of understanding for the user that each conversation is built off past conversations, growing the relationship between user and voice experience.

Conversational markers:
Use words and phrases like “thank you”, “got it”, and “please” to personalize the conversation and speak to the user in a way they’re used to from conversations with other humans.

Here are some examples of how RAIN helped their clients solve for this:

• RAIN worked with Sesame Street to recreate Elmo’s voice in a kid-targeted voice experience, ensuring they used the appropriate language, voice, and tonal adjustments to ensure an authentic Elmo and Sesame Street feel. RAIN took thousands of clips and strung them together to form scripted Elmo statements and responses, accounting for a variety of scenarios in which Elmo may have to react.
• For Starbucks, in order to incorporate within the Alexa Skill some of the in-person characteristics of a barista that make the ordering experience successful and unique, the RAIN team visited brick and mortar locations to listen to how the baristas spoke. This helped them craft the voice and language of the experience, understand the importance of using a customer’s name, and more

User Research and Testing Advice

Dailey of RAIN shares a couple of ways to test the persona, voice, tonal adjustments, and anthropomorphic characteristics you’ve selected for your voice-based experience. “The simplest way to fix something that can occur early on in your voice design process is a targeted survey to, or focus group with, identified audiences.” She adds that a survey can help gauge immediate reactions as well as can uncover unexpected insights for your consideration prior to actually designing and developing a prototype.

“Another way to test these elements is through user-testing of a low-fidelity prototype.” Having users log their interactions and reactions to those interactions can help you get a sense of how well the experience is providing value and connecting on an emotional level, if at all.

In our next chapter, learn more about infusing personality and humor into your VUI. With the right amount of wit, banter, and ease, your brand voice can be much more human.