The new “Voice” interface requires a new way of thinking in many aspects – not only in content creation, but also in the conception of user experiences and interface design.
Alexa developers of the moment – and not only us – are doing pioneering work. We remember the beginnings of smartphone app development around 10 years ago and we can still see today that many things are being implemented for the very first time. No stackoverflow tickets, no best practices, no apps that offer inspiration help. Bury your head in the sand or face the problem? We have opted for the second choice!
With QuizClash, our voice assistant agency 169 Labs has developed the first multiplayer game for Amazon Alexa in Germany. Users can seamlessly switch between the mobile app and the Alexa skill. Often developers for Alexa have dreamed that users start an interaction on a smartphone with a fingertip and continue it with speech – the term multimodality is the corresponding synonym.
With the famous “Casual Game” QuizClash we have dared to approach this dream and make it possible for the 36 million users in Germany to challenge their friends by voice. In other projects we had already learnt that content had to be speech-compatible and could not simply be transferred 1:1 from the web or a mobile app. We often give this advice in our workshops, in which we teach the basics of UX conception for applications for language assistants. We simply have to understand that voice is not just another channel, but a whole new ecosystem. It requires new workflows to make content and concepts interoperable.
In implementing the QuizClash skill, we encountered the natural limits of the transferability of screen/touch UX to voice.
The motto was: The visible becomes invisible, but must not go unnoticed.
During the development of QuizClash, we intensively studied the game concept we had learnt and determined which of it could be implemented meaningfully via language and at which points we had to make adjustments. Of course, we still had to remain true to the principle of the game, so that users could play it equally via the mobile app and the Alexa Skill.
Here are our 5 biggest learnings from the implementation of the QuizClash skill:
1) Transparency over choices
Although there are now Echo devices with displays, such as the Amazon Echo Show or Echo Spot, about 90% of the devices on the market are without a screen. The golden rule is: Voice first. This poses a great challenge to developers of voice apps: without embarrassing or overlooking the user, we have to serve him as many decisions and choices as possible, in the truest sense of the word, “bite-sized”. That led to the fact that with QuizClash on Alexa the functional range had to be limited to the basic functions, since it is nearly impossible without visual anchors from user view to keep the overview. After successfully linking their account, users can currently start new games against any players and friends or continue running games. This sounds quite simple at first – but it was complex enough to require several thousand lines of code. When starting the skill, about 10 case distinctions happen in the background, which affect what Alexa says to the user next to a “hello back!” The crucial thing is: Do I have one or more running games? Is it my turn or not and if so, how many games and against whom? Do I have one or more open challenges? Who challenges me? Do I want to start a new game against any player or friend? If it’s a friend, who on my list? Who can I start a new game against at all?
All this information (and more) influences what Alexa says to the user in the welcome message – and what not. On a smartphone screen, we simply display all the information next to and below each other – the user will find out what he needs or wants to use.
The dynamic reloading or buffering of data must be handled in milliseconds in the voice context. The user gives a right or wrong answer, the skill then communicates with an API and must immediately advertise the next interaction. In the mobile context we can insert a “Close” or “Back” button and show the following view only when all data is fully loaded (e.g. with lazy loading). On Voice, each close, exit or continue interaction is another question that costs the user several seconds in case of doubt and thus allows a natural frustration limit to be reached more quickly.
3) Input or output of player names and answer options
In QuizClash, there is no convention for naming a player’s name. Special characters, emojis, numbers – anything is allowed. The questions occasionally contain technical terms or words from another language. Alexa’s pronunciation of these words is partly not natural yet, the input of these terms is hardly possible from the user’s point of view. For this reason the voice input of user names was not possible and during the quiz the users can only answer with A B C or D during the duel and not return the word from the possible answers. This could change in a subsequent version of the skill, but the quality assurance of approx. 27,000 questions is correspondingly complex.
4) Temporarily inactive functions or selection options
UX concepts for screens use .active, .hover, .visited, .disabled or other classes to show users if and what they can and cannot do with this choice. On Voice we have to implement all these helpful features invisibly, without frustrating the user. A negative example would be: At each step, explain in detail which choices the user currently has and which (temporarily) do not. The friends list in Skill, which you can use to send new challenges for a quiz round, is a good example of this. On a screen, I just gray out all the users I can’t challenge. Do I just omit these names in the voice context? What if I can’t challenge any of my friends at the moment because I already have a game running against them all? What if I don’t even remember the cryptic usernames of my friends? The biggest “challenge” was to make these and other things user-friendly, especially when implementing the Alexa skill.
5) Playing without a user account
QuizClash has not yet provided an option for users who do not have an account. This account is created within seconds on the screen. Our claim was to make it potentially possible for every Alexa user to use QuizClash. This resulted in a completely new game concept: In “training mode”, players can answer 10 new questions a day and find out how well they beat the rest of the players at the end of the day. This is an important extension to allow users to play a quiz round without any media discontinuity (entering the user name and password via the Alexa Companion app).
Let’s keep in mind: Voice is a new ecosystem and requires a new way of thinking and a lot of creativity.
We learn from user feedback and improve UX from day to day. The launch in England and the USA is imminent, the QuizClash app is currently available in 16 languages.
The development of QuizClash for Amazon Alexa was pioneering work: Users can now play the popular quiz game against other players hands-free in the kitchen, on the couch, in the bathroom or in the car in addition to their smartphone. All they have to say is: “Alexa, starte ein Quizduell!” (The skill is only avaiable in German for now)
After the ALL ABOUT VOICE Conference 2018 was met with an overwhelmingly positive response, we are now presenting...READ MORE
A well designed testing process is vital for developing successful voice applications. The benefit of a proper process...READ MORE
Amazon anncounced the Alexa Skills Challenge for developers in the United Kingdom, Ireland, Germany, and Austria: The Challenge...READ MORE