Designing a mobile car buying experience
Duration: 8 months
Team: UX Designer, Product Manager, Data Analyst, Frontend/Backend Engineers
-
Car auctions move fast, and Byrider needs to sell approximately 1.2 million cars a year. The Indianapolis-based company approached my team to help build a mobile app to track purchases and help field buyers communicate with their home office while they were at the auctions.
At the time of engagement, they used spreadsheet printouts that listed potential vehicles to bid on. These spreadsheets were many pages long, and had the possibility to quickly become outdated since the auctions digitally updated their inventory up to the minute.
This process caused field buyers to miss out on important purchases, since they often had a large amount of cars to assess and bid. If they won, they needed to report the purchase within minutes so that the home office could add it to the inventory list.
-
I worked with a cross-functional team to conduct initial discovery with stakeholders at Byrider headquarters. This discovery consisted of diary activities, contextual inquiry and mapping the field buyer’s journey. After initial research, we gathered enough assumptions to begin the design process.
I worked in grayscale, with wireframes setting up the app’s architectural navigation. I used hand sketches to convey early ideas to product and engineering teams. Engineers used my sketches to build an MVP in React Native. I conducted visual QA throughout the process, and incrementally collaborated on UI refinements and interactions.
I held weekly validation sessions with users and stakeholders. I planned mini activities for each week’s session to helps refine the problem together. I interviewed participants to gather their needs, shared work in progress, as well as engineering insights.
Once the team felt confident to test concepts, we visited Byrider field buyers in Indianapolis to run usability sessions with auction buyers. We used a pre-production build to collect early data from field buyers. This helped inform the next phase of the MVP.
-
Using the data model proved successful for Byrider employees. This initial phase helped us map out next steps for the product, like promoting suggested vehicles to bid on within the app based on collected data from the field buyers using the MVP. We mapped out ways to record daily progress at the auctions, such as inputting an estimated bid price, and then updating the actual purchase price when finished. This would help gather more accurate data, while also reducing the burden on home office admins to manually keep track of this information.
We also made recommendations to track and share notable auctions in their area. This way, teams will have the ability to coordinate their efforts and reach more events–resulting in adding more cars to their inventory on a given day.
Initial flow diagram
-
Documented tasks within the flow. Used sticky notes to iterate on task order before finalizing.
Initial wireframes
-
Worked in grayscale to refine screen flow and interactions. Began brand exploration alongside the wireframes.
Initial prototype
-
Used Adobe XD to finalize visual look and feel. Worked with engineering to build a testing prototype in React Native.
Building a Census Mobile Toolkit
Duration: 1 year
Team: Interaction Designer, Mobile Engineer
-
The Census Bureau constructs various types of survey methods to collect different kinds of population data. They can ask field responders to assist with data collection, or solicit people to self-respond through either web or paper-based surveys. These surveys have different layouts so that they can match the response type.
Up until recently, Census relied upon paper surveys or field responders knocking on doors. Census Innovation Lab researchers designed studies to inform how survey responders might use mobile devices or native apps to respond to their surveys.
-
The Census researchers designed usability studies to inform how study participants used mobile devices to respond to different types of questionnaires. These findings were designed to help shape native mobile guidelines specific to mobile-based surveys Census. This would also help disseminate standards and create continuity across the different survey types.
I worked with the researchers and mobile engineers to design and build a number of native mobile prototypes and production builds so that they could conduct usability tests with participants. We used the American Community Survey to influence screen elements, and generate potential patterns to test. The researchers recruited assisted living residents (age 65 and older) who used mobile devices, as the demographic closely mirrored their existing survey participants.
-
I used Apple Human Interface Guidelines and Android Developer guidelines as a foundation to determine a baseline for successful native mobile standards, and help inform what else we might not know about mobile usability. I consulted with the researchers on visual layouts for the studies, and how to validate device-independent patterns considered in the research.
I worked with the engineer to embed specific metrics to help track usage during the study. We tested the performance of different components like button and input field styles. I created a range of different patterns so that we could gather both qualitative and quantitative feedback, like user satisfaction, conversion and task completion time. The study design approached designing for trust, which I reflected in the diversity of prototypes. For instance, some included the Census branding and were heavily designed, while others were more reductive. We used TestFlight builds to run the experiments on iOS devices.
-
The findings that the team collected informed recommendations for Census survey patterns. We outlined ways that these patterns may contribute to a design system, which would align departments across the bureau. We defined ways to leverage trust through these patterns, and build upon native device settings to optimize the experience. We helped the innovation team document the findings, and drafted plans for future rounds of testing with other study cohorts, like field responders and volunteers.
Touch target
-
Determining optimal spacing for touch targets to inform the guidelines.
Screen elements
-
Exploring form input components that invoke clarity, trust and delight.
Hypothesis was that emotive content would sustain engagement for longer surveys.