Do you understand me?
Voice User Interface tips using Nielsen Usability Heuristics
I’m working in Native Voice User Interface for a senior medical wearable, and my persona is a 75+ year old with very little technology experience or patience. I adore my user group. Voice and Conversational User Interface can be challenging because there are so many nuances in conversation.
Communicating with a hearing-impaired 85 year old can sometimes be complicated, and more so when that conversation is through a device. The conversation can also be full of laughs. We may not always be on the same page of what we’re talking about, then add the fact that a hearing aid is either picking up too much or too little background noise, can also lead to miscommunication, joy, and frustration.
Research sttes that by the age of 70, most of us will be facing a hearing and/or visual impairment caused by an age-related disease. The oldest of baby boomers are 70ish now and so this is a generation that can benefit from an emerging Voice User Interface (VUI, pronounced“Voo-E”) technology. Designing a conversation in Voice UI for a senior requires complex thinking of the dialogue flow, yet fun and rewarding to create. I am a purpose-based UX designer and threw myself head first and heart first into my native VUI project.
When designing the dialogue flow for our persona, the Silent Generation, I chose to apply Jakob Nielsen’s UI usability heuristics for guidance. The ten usability heuristics were developed in the 90s as guidelines for interaction design. Conversation is a constant interaction between people so I thought it would be fitting to apply the principles to my VUI project where a person and their wearable were communicating and interacting. I also attended Amazon Alexa and Google Home workshops as they have VUI design outlines, but we aren’t using Alexa or Google, so I didn’t solely use their ideas and guidelines. I definitely did not want an Amazon Echo Silver to become reality. Allegra..Alexandra..Uh huh…
I decided on the tried-and-true heuristics and applied them to my project. The 10 Heuristics for User Interface Design are ideal when applying to Voice UI. It’s no surprise since Nielsen and Norman are THE usability kings. They know what they’re doing and there’s a reason why these guidelines have stood the time over the past decades. These heuristics are stellar. Here are the key points in Nielsen heuristics that are most useful in my voice design process.
Visibility of system status
- Write dialogue scripts so that users know their progress within the conversation
- Take advantage of sound and visual capabilities and use as signifiers to give regular user feedback
- Make interface and commands discoverable so your user knows what to say and when to speak
Match between system and real world
- Use talking words that are in a conversation; think telephone vs texting
- Design for emotion; give the voice a personality that’s likable and congenial
- Sentences should feel natural and spontaneous and, artificial syntax and non-conversational words should be avoided
User control and freedom
- Allow user to easily ask the device to repeat questions or commands
- Ensure user’s ability to exit conversation
- Respect your user and don’t allow the device to interrupt him or her
Consistency and standards
- Identify conversational patterns with your persona and user group
- Allow user to not have to think hard about what to say — if you have commands, keep them simple and easily memorable
- Know your persona and their listening, hearing and understanding capabilities (e.g., seniors understand low pitch sounds more than high pitch)
- User should not feel at blame for the device not understanding them
- Focus on preventing errors as the voice device’s purpose is to serve and support the user’s needs and wishes
- Give users positive reinforcement with a confirmation, but be careful to not be didactic or repetitive
Flexibility and efficiency of use
- Context is everything in a conversation interface; consider colloquial synonyms in script
- Set user expectations; don’t assume that user knows what to do
- Research and understand how questions are asked, and answered
Aesthetic and minimalist design
- Simplicity. Less is more and keep sentences brief
- Avoid cognitive overload by clearly presenting options and not too many questions or options
- Consider grammar and syntax; cut out extra words like qualifiers such as adverbs, passive voice, and jargon or slang
Help users recognize, diagnose, and recover from errors
- Make it clear when users need to respond
- Don’t expose users directly to error handling
- Let user know what and where the error is and direct them to a place where they don’t make an error. This takes work.
Help and documentation
- Offer help for complex situations
- Empty pauses (time) make user feel like they are doing something wrong
- Best if voice responds immediately if user needs help
Voice UI is tricky and worth the effort
Voice UI can be tricky and easy to get it wrong. Everyone hears, listens and understands differently, and all of it takes brain-work and concentration. I think it’s useful to observe that this challenge exists with all UI work — but with voice there are fewer cues and signals — less context, so the stakes are higher. Even at the dinner table, someone could have a conversation with another person and what one person understands is not what the other person thought they said. Miscommunication is everywhere, which is not the intent of voice interfaces. It’s definitely not the intent of UX voice interface designers. We’re trying to design devices that help with efficient, accessible hands-free communication, not create more miscommunication.
If it’s a badly designed website, then the user just bumbles through in hopes of finding a valuable morsel of information. If it’s a badly designed voice interface, then the device will probably be turned off and we all lose out. Nobody wants that, most of all the user. Users want to be successful and businesses want them to be successful too. Particularly in my case where the wearable is a medical device designed for our users’ personal health and safety. Continued user testing and iterations, before and after market launch, will be key in solving any miscommunication between the voice interface and our senior users.
Currently, the Voice UI revolution is trickling down from the big corporations such as— Amazon, Apple, Google, Microsoft, IBM, etc. This is different scenario than software creation over the past two decades where ground-level companies and start-ups were driving the innovation bus. I am thrilled to be part of a Native VUI project that is working on a creative platform. Thank you Nielsen Norman Group for your steadfast wisdom. We need more competitive native VUI platforms. Competition is the avenue of innovation and creation.