Mozilla Firefox Listen
"Creating a voice browsing future"
Problem Space
In 2017, the U.S. Census Bureau reported that the national average commute time is 26.9 minutes (one-way). For some of us, we are always on the go. One interesting area to explore is voice control. How can we build smarter products that leverage the power of voice control to make our life easier?
Project Overview
We partnered with Mozilla’s Emerging Technology Lab to conduct a usability study and see how we can improve their upcoming audio streaming app, specifically the hands-free feature.
Challenges
The challenge for this project is to work with existing designs and to build a prototype that supports voice interactions.
Role
Usability Testing
Duration
3 months
Tools
Sketch
Principle
Invision
Team
Kent Tsai
Christina Zheng
Haihua Zhang
Siddhant Patil
1 | Usability Study Goals
Understand user’s impression and experience with hands-free feature.
Understand user’s initial understanding of the unique playlist structure.
2 | Define Users
In oder to understand how users felt about the features, we recruited 9 commuters in Seattle. These commuters either travel by car or public transportation and have experience using podcast apps.
3 | Preparation
Usability Test Kit
Screener questionnaire
Consent form
Interview script
Pre-test questionnaire
Post-test questionnaire
Data logging form
Technology
Camera (Record setting)
GoPro (Record screen)
Tripod
Invision
Phone
4 | Research Method
In order to understand users’ experience with the app, we chose 3 methods for our study. First, users will give us a cognitive walkthrough of what they think the UI represents and how they expect different elements to behave. Next, users will think out loud as they go through the tasks and elaborate their thoughts in more detail. Next, we choose Wizard of Oz as a method to simulate the voice interaction.
Cognitive Walkthrough
Think Out Loud
Wizard of Oz
5 | Data Analysis
Affinity Diagram
In order to organize and analyze the data we collected, we completed an affinity diagram exercise. First, our group members are randomly assigned to two participants per person and we coded the participant data independently. During the coding phase, we identified key highlights and surprises that stood out to us from our field notes and the video footages. Then, we used Realtimeboard (Miro) to create virtual Post-It notes with our key highlights. Collectively, we reviewed all of our notes together and identified common themes and surprises for further discussion.
Step 1: Encoding
Step 2: Grouping
Task Analysis
We also conducted a task-analysis to understand the participants’ experience through completing each task. First, the moderator introduces the tasks to the participant. After the participant tries to explore and identify ways to complete said tasks, the note-taker will record whether or not the participant completed the task successfully. There are three levels of task completion: successful, partially successful, and fail.
6 | Key Findings
Most participants expect to use voice for turning off hands-free mode.
Most participants cannot distinguish between playlist items due to lack of differentiating characteristics.
Some participants are confused by the hands-free icon, lacking affordance.
7 | Key takeaways
People feel uneasy using voice control in public, especially in public transportation.
It is crucial to have a clear feedback signal in any voice interaction to let the user know that their input have been registered and processed.