Identify a problem:
Due to the stresses in people's lives, especially during the times of a global pandemic, people are finding it harder to sleep.
So, a solution was needed where people can speak to their voice assistant to calm down, relax and enjoy a better sleep.
Whimsical, Voiceflow, Logic Pro X, Amazon Developer Console, Dashbot.
After identifying the problem, conducting both primary and secondary research was incredibly useful to analyze and further understand the problem and pain-points from the stakeholders/audience perspective.
Through the secondary research I was able to understand the wider sleep problem, gather some hard data and back up the project that answered scientific questions such as:
"What effect does sleep have on the body?”
By conducting user research interviews, I gained a further understanding from people who have had, or currently have problems sleeping. Through this method of research, I was able to connect with potential users on an emotional level, finding out what their struggles were and how they currently try to eradicate the problem.
I then produced a competitive analysis, to identify the current market and what areas were lacking and could be identified.
Target Audience & Psychographics
By starting with a broad range of features, I was able to narrow them down and establish a project timeline: now, near and far, to define and prioritise the most important features. By utilising the 4/6 Thinking Hat method, it allowed me to assess each feature in order to help with the decision making. Through this, I was able to take a holistic view, by looking at each feature with a different set of eyes to identify any concerns or emotions that could affect each feature.
It was important to research, identify and visualise the target audience through demographics and psychographics to gain a better understanding of what drives and influences them, therefor keeping the user-centric approach running throughout the design.
Crafted a user persona from the research (customer profile) to establish the type of person the project was aiming for.
Created a customer mood board to help with the visualisation and help when sharing project with the team or stakeholders.
I spent time to define the assistant persona, to define what personality the voice assistant will have and how it will act with the user. By keeping the user in mind, I was able to use the key words that were mentioned by users in the user research, to establish the personality.
Trusting, Supportive, Calm, Attentive, Warming.
Whilst defining the persona, I also identified the functions and limitations of the assistant that were established during the features phase. This helped to further cement what the assistant could and couldn’t do.
Click images to enlarge
The initial phase of the design process started with the creating the Conversational Components.
These helped to define the standard vocabulary of the assistant that resonated the personality, whilst keeping a solid consistency throughout each stage and scenario of the design.
- Introductions, confirmations, acknowledgements, error handling and endings.
Sample dialogs enabled me to draft out the conversations throughout each use case and scenario, such as first-time user vs returning users, happy path vs error or edge cases. Here I used a very simple method of the Wizard-of-Oz technique, to test and improve the copywriting and to find out what extra paths I needed to design.
Click images to enlarge
In order to better understand how the wording and speech would sound, I used Google’s online TTS engine to play out the dialog. I was then able to alter various aspects of the dialog to make it sound more natural for speech. The examples below show the very first versions of welcome message.
Intro - Demo Version 1
Welcome to Imagine Sleep Sounds. I can help to relax your
mind and fall asleep by playing breathing exercises and
ambient sounds. Which would you like to hear?
Intro - Demo Version 2 - (after listening on TTS engine)
Welcome to Imagine Sleep Sounds. I'm here to help you to
fall asleep and relax your mind. I have a number of breathing
exercises and ambient sounds for you. Which would you like
Following that, it was important to produce a flow diagram of the entire conversational flow to understand the overall structure of the design. Only once all the previous steps had been completed, I began the design using Voiceflow prototyping tool.
Imagine Sounds Flow Diagram
The natural sounds that were chosen were: Ocean, Rain, Birds and Jungle/Rainforest. These were chosen due to the research stating that these natural sounds are popular for relaxing and for improving sleep. I was able to use my sound design skills to create and mix different sounds from my own personal sound library that I have either recorded myself or collected over the years.
I sound designed and created the earcons, aiming to resonate with the calm and warming assistant persona. Different versions of the earcons were created to provide subtle changes depending on the user journey and intent, but also ensuring a consistent theme was kept throughout the design.
Google's Conversation Design Guidelines
Earcons, or audio icons, are a type of non-verbal audio (NVA) designed to convey a specific meaning. Your persona can use earcons to communicate with users in the briefest way possible.
TESTING AND ITERATIONS
New vs Returning Users
Once the first version of the design was complete, I conducted usability tests to identify any issues and highlight any areas of improvement. In the usability test report, I structured the problems and errors with a severity scale ranging from minor to critical, along with further analysis and recommendations. This helped to work out which needed the most urgent attention.
One of the main user concerns showed that it took too long for the user to get to the end goal. A difficulty I had within this project was how to explain the information needed, in smaller pieces with time efficiently a priority.
I was able to solve this problem by streamlining the process and utilising different paths and prompts for new vs returning users. I reduced the number of steps from 3 to 2, by changing the style of the copywriting and prompts to not provide all the information at once but by allowing the user to choose and decide their path.
By looking at the usability testing and comments from the users, it became clear that I needed to alter how the responses were structured for new vs returning users. By increasing the user session number every time the user opens the skill, I was able to create separate welcome responses depending on how many times they used the skill, therefor creating a much more personal experience. Each welcome message contained several variations to keep them varied and interesting.
Other improvements included changing the synthetic voice from a British male voice to a more flowing and natural US male voice, whilst also changing the names of the breathing exercises to ease confusion amongst the users.
The full usability test report can be viewed in the PDF link.
Video: Step Reduction
Screenshot Voiceflow: New vs Returning User Variations
Data, Analytics and Looking Ahead
With the design compete and ready for deployment, I implemented Dashbot's API's into the Voiceflow project, at each stage that I wanted to collect data. This data will be analysed periodically, to identify common issues or areas of improvement. Further design iterations and adaptations will be continually processed to ensure the success of the skill.