Overzicht Actueel

Designing voice user interfaces (3): knowing your voice brand and responses

In the previous article in this series about designing voice user interfaces, we explained the importance of persona design before starting the voice project. How important it is for your brand to come across (just as you intend it to) and it enhances the complete user experience. As Cohen, Giangola and Balogh said in their book Voice Interface Design, “There is no such thing as a voice user interface with no personality.” If you do not give it a persona, your customers will. Now, how can we apply and strengthen the persona? By looking at the way consumers communicate with voice user interfaces and what responses to give back.

Embrace the essence of your brand
Let’s first take a step back. What is the reason consumers do need that persona? Pareidolia: people have the tendency to see human characteristics in objects. Just like my neighbor named his car Ronaldo. Given the fact that a voice interface talks back to the user, they cannot help but see human characteristics in it.

Google had a big task at hand to be able to design personas for every country’s Google Assistant. Benjamin Dorvel who is the Staff Conceptual Designer, Human-Centric Strategy & Innovation at Google and he manages this whole process. He explained it to me during a Google Hangout of about an hour. The Google Assistant at its very essence has a couple of characteristic values. Values such as, empathic, decisive, playful, social, optimistic, and friendly. In order to deliver a consistent brand image. These core characteristics are not up for discussion. They exist in every Google Assistant across the world. Each Google Assistant in every country has the same foundation, yet it is slightly adapted to make it more fitting for the country in question. Google uses creative writers to add other characteristics for the persona in order to make it match the country. One country specific characteristic is humor, that we all used is “Hey Google, tell me a joke”. A joke in the Netherlands would not be the best joke in the U.K.

Just like Google is taking notice to add elements to its core characteristics, your company should do so too. What does your company stand for and how does it want to be perceived? When making several Google Actions do you want to have a core that is the same across them all or should they be vastly different? By adding your organisation’s core characteristics to the Google Action it will enhance the brand experience and consumers will have a better consumer experience since they have similar characteristics. They have the feeling that they get to understand your brand better. What if your brand was a human being, what kind of personality does it have? How do you want it to come across to your audience? Think about a major media organisation that might have radio, television, magazines, etc. At the core all parts of the organisation want to be open, approachable, likelable, social, emphatic. Yet one part might be more aimed at teenagers so it would alter the way they talk to the user, while keeping those core characteristics intact.

We talk the same across cultures
Most people believe that people in different cultures also speak in a different manner. But is that really the case? Elizabeth Stokoe, author of the book Talk the Science of Conversation proves otherwise, the way people talk among various cultures is not that different at all. People everywhere make requests, offers, invites, complaints and the list goes on. Sometimes we do have myriad resources to get them done. However, these actions are similar in all human cultures, such as the way a Chinese versus an American person would turn on their favourite song on a voice device. The majority of these actions happen via talking. What is different between cultures is the linguistics and multimodel resources people have at their disposal. In the linguistics there are context, the contingencies, our obligations and our entitlements that make the difference.

Two key design aspects for voice are the context the users are in and the contingencies. Depending on the context of the user, the responses they get should be adjusted accordingly. For instance, when us users who talk with a smart assistant at home may have more time to spare as compared to a user on the go running between two appointments. Or when the user is cooking in the kitchen as compared to talking to your voice assistant while driving. Again, very different contexts to take into account. Where is your user while making use of voice devices and what kind of voice devices do they use? Make an outline of the customer journey. In the customer journey it needs to be clear where the user is, bathroom, kitchen, in the car etc. How much time are they spending in that environment, do they actively pay attention to what the voice assistant says and during what time of the day is the user in that space? Is the voice assistant user alone or together with a group of friends? If you have mapped all these factors and know during which x amount of time of the day your consumers use the voice assistant you can anticipate how the user communicates with the assistant and how your brand should interact with the user in that moment.

Even when knowing the customer journey, knowing everything the user says is almost impossible. Contingencies, unexpected turns, are bound to happen and these are the most valuable part of the conversation. Why? Just imagine that you are in the supermarket and want to know where the pineapple cans are. The person doesn’t understand the question so he continually replies, ‘sorry I don’t understand that’. Needless to say that conversation ended. You do not want to lose your user, in most cases the user does not have a screen to fall back upon to make sure the users can do something with your response. By adding the right amount of speech in your response you will help the user understand what they can do and how they could do it. Each time an user is not understood, add more detail to it. Let the user know what your Google Action can do, this is called the escalating detail. In case after two attempts the assistant still does not get the user and you have a lot to offer, send them to your mobile app. Provide them with a link where they can press upon and educate them about what they can do in the Google Action. Depending on the context where the contingency took place you can provide the user with a voice only escalating detail or send them towards your mobile app to help them further along there.

Focus on your foundation first, don’t let your user get lost
When you know the core of your brand and how you want to transfer that into voice, you know how to handle the user actions in their context and know how you want to deal with contingencies. In that case you are well underway to finalize step 1 out of the three conversational stage, a model made by Soundhound.

Source: Presentation Soundhound during Voice of The Car Summit, 8 April, 2020

The three steps show the importance of making sure you deliver what your consumers are expecting from you. At first, it might seem easy yet many do forget to deliver. The amount of times a Google Action or Alexa Skill said, ‘Sorry I did not understand that’ is too high. Make sure your core functionality works and make sure you know what you will say when something goes wrong.

After this, you can go on and think about how you want to set your voice application apart from others and enhance the voice UX by understanding the follow-up questions. Finally, if the first two steps work and the consumers are using your voice app. Then you are ready to add delightfulness to it. Think about the Google Action from Walt Disney. They first made sure the Google Action worked when the user talks to it and gives answers at the right speed. Then they differentiated by making sure that when the user is reading one of the selected Disney books the Google Action exactly knows where the user is. Even when the user stops reading and continues the next day. To delight the user they added sound effects to certain parts of the book. By carefully following the three steps Disney was able to create a unique Google Action that definitely delighted the users and brought the Disney magic into your living room.

On May 14th, we organise a Digital Talk about how to create the best user experience for voice apps, with great cases from companies like bol.com. Feel free to join us (the event is in Dutch)!

Up next: Conversational Flow
So far you have learned:

  • The basics of voice design (define your goal, define your target audience and to define the scope).
  • You’re also aware of the importance of crafting a proper persona which fits your brand.
  • In this article, we discussed the essence of your voice app, why context and contingencies are important, and to make sure you get the basics right first before extending or differentiate yourself and delight the user.

To make sure the conversation with your consumers goes according to plan and to know how to respond with the right device options, next time I will give insights in the creation of the Conversational Flow. To make sure the sentence, ‘I am sorry I do not understand this’ never appears into your voice app.

Stay tuned!

cijfer voice ux