.png)
Domino's Dom Bot Remote User Testing and Redesign
User research and recommendations for Domino's Dom Bot, the dominos.com AI help chat bot
Project Scope
Domino's is a staple in weekend eats. Their online ordering platform is found at Dominos.com. They recently added Dom Bot, an AI help chat option that offers a simpler ordering option. This accessible wizard is only helpful if people can find it. The goal of this research was to uncover pain points and areas of improvement for the Domino's Dom Bot.
Issues we would like to investigate with the interface on dominos.com include their AI assistant ‘Dom the Pizza Bot.’ Users may not know the Dom Bot exists or know how to access it. We were focused on the question:
When is the ‘Dom the Pizza Bot’ being used, and where does it add value?
​
​
HIGH LEVEL TIMELINE
1 week to explore the website and develop test plan, 1 week to conduct remote usability testing, 2 week to analyze and synthesize data and develop 3 recommendations, 1 week to redesign
THE RESEARCH TEAM
Gillian Wasserman, Jess Rakowski, Aroon Prabhu, Tanvi Hedge, and Nathan Gupta
TOOLS
SKILLS
UserTesting.com, Microsoft Excel
Remote user testing, data synthesis and analysis
Methods
We conducted a remote usability test of Dominos.com via Usertesting.com. The tasks for our users were created to allow us to observe typical usage of the Dom Bot and 2 follow-up questions to understand the thoughts and feelings of the user post-test.
The Tasks:
​
Task 1: Place a carryout (Zip Code: 48104) order of a large two-topping pizza with the following arrangement: cheese everywhere, red sauce, mushrooms on one half, pepperoni everywhere, and garlic crust.
​
Task 2: Use Dom, the virtual assistant, to place a carryout (Zip Code: 48104) order of a large two-topping pizza with the following arrangement: cheese everywhere, red sauce, mushrooms on one half, pepperoni everywhere, and garlic crust.
​
Post Test Questions:
​
Q1: What did you think about using Dom, the virtual assistant? What made you feel that way? [Text response / Audio response]
​
Q2: How likely are you to use Dom, the virtual assistant, again? [Likert scale]
Data Analysis
The sample
-18 participants
-under 30 years old
-in the United States
-average web expertise
​
Task 2 User Journey
The data
-18 screen recordings and real-time voice overs of each participant
-answers to post-test questions
66% of user test participants struggled to find Dom
22% of user test participants left task #2 incomplete because they were unable to locate Dom
Valuable quotes
“The virtual assistant helped more than I thought it would”
​
"I don't know how to find the virtual assistant"
​
Dom is difficult to find
To our surprise, users were unable to find the Dom Bot virtual assistant before they could even use it. Since a large group of our sample was unable to find it, we recommend repositioning, resizing, and reshaping the Dom Bot access-point so that users know where to find it.
Those that were able to find Dom and use it enjoyed their experience
The users that found Dom thought it was a useful tool on Dominos.com. Since this feature holds value for some customers, it is important for users to be able to find it. This contributes to our recommendation of making the Dom Bot easily findable.
Key Findings and Recommendations
Redesign
The Original
Text: "Order with Dom"
Font Size: 14 pt font
Button Size: 146px x 38.5px
Positioning: floating button
Location: bottom left corner of browser window
Redesign
Text: "Need Help? Order with Dom"
Font Size: 18 pt font
Button Size: 139px x 95px
Positioning: tethered to edge of browser window
Location: bottom right area of window
Key Considerations
Visibility and Recognition
Since users were missing the Dom Bot button on the home page, I started by changing the text on the button. I changed "Order with Dom" because that text does not resonate with users that aren't familiar with Dom. I reworded it to "Need Help? Order With Dom" to catch the attention of users with a question and to provide context.
I also aimed to make the button more visible. To do so, I increased the font size by 34% and the button surface area by 135%.
​
I adjusted the positioning and location so it is in a place the user expects to find a help button. The bottom right corner is the typical place to find a help button, so I brought it there. I also tethered the button to the side of the screen. This is crucial because the original button's positioning overlaps with the landing page/catalog screen and gets lost in the content when the screen size changes. By keeping it stuck to the side of the window, the button's location is static when the screen size changes.
​
​
TITLE OF THE CALLOUT BLOCK
LESSONS LEARNED
User testing will not usually give results that you would have never anticipated. Embrace the unexpected.
​
Good products and features deserve to be noticed and used. If the user can't find it, lead them there.
​
​