What did I do at Uber?
During my summer internship and as one of Uber's first Innovation Lab designers, my main project included designing a new administrative system for our new driver onboarding platform.
I was tasked with designing a dashboard which would quickly and seamlessly onboard new Uber drivers in person, live with an Uber representative at a kiosk.
Previously, onboarding drivers was possible only through the app and isolated large demographics of potential drivers who weren’t comfortable submitting their private data online.
Our kiosk allowed new drivers to ask questions directly to a person, to finish and submit their required paperwork, and to manage any other problems they had through a person-to-person interaction. This gave the new drivers a sense of confidence and security in their new pursuit.
tldr:
One of the biggest challenges I faced throughout the design process was selecting the proper datasets to visualize and display. In order to make an impactful product, I needed to surface the most important information that was previously tracked in Excel spreadsheets or that had never been documented before. Choosing the proper data visualizations was a critical component for this dashboard as I wanted to design a system where users could easily glean the insights they needed.
The version you see below is the product of several iterations and I’m proud to say that the kiosk has been developed and launched in several Bay Area malls and in the Toronto International Airport.
Our ambition for the future: requesting a ride, wherever you are, without your phone. (photo via: https://eng.uber.com/uber-kiosk/)
Process
This dashboard was designed for internal users to manage and track analytics related to new driver onboarding initiatives. I was designing a new product that had never existed before and in order to create something useful, I had to understand the user’s needs and problems first.
My iterative processes included:
Conducting user research and ethnographic studies on our end user
Holding weekly meetings with each stakeholder to ensure needs were being met
Running an iterative and human-centered design process to ensure we were solving the right problem
Creating a product roadmap to share with the team
After synthesizing this information, I began sketching and creating basic wireframes.
Once the constraints were outlined and the functionalities finalized, I moved over to Sketch where I began to bring these ideas to life in higher fidelity prototypes. Throughout this process, I maintained constant communication with the product and engineering teams to keep everyone up-to-date on the status of the project and to troubleshoot any issues. I also used this time to improve the overall UX and UI by eliminating extra clicks, creating a more concise site navigation, and by reducing the overall amount of content on each page. I then moved on to bring in Uber’s sleek, modern, and lightweight brand into the product.
The Uber Kiosk which launched in 2017 (photo via: https://eng.uber.com/uber-kiosk/)
Aspirations for the next iteration of the Kiosk (photo via: https://eng.uber.com/uber-kiosk/)
Results
All kiosk related data was now organized & centralized in one location
Crossed 1000 interactions across 13 malls in ~3 months
Kiosk launched to 20 additional sites by end of 2017
The kiosk was presented to Uber’s CTO
Customer satisfaction rose from 81% to 90%
Net Promoter Score went from 61 to 93
Expert helpfulness rose from 4.48 to 4.58
“The biggest test for a product is to test its resiliency when it hits reality.”
Challenges
As we all know, the design process is never straight-forward nor does it come without its challenges. For this particular design, there were no data-related problems like data latency or data pipelines failures. Instead, the challenges stemmed from existing datasets which lived in Excel spreadsheets and from collaborating across several teams to ensure that value could be derived from this product for each team. Because the datasets either weren't being tracked or were updated slowly over time in static documents, I had to find and ensure that the most valuable information was surfaced in this dashboard. What's more, because the teams involved in this project were spread across the country, we had to overcome collaboration barriers like the transfer of information to correct parties. What resolved most of these issues was stepping back and properly diagnosing the challenges.
At the start of the project, I noticed that metric data was being collected and that it was being organized methodically and efficiently. This was a great thing. However, once this was done, the data wasn't being shared with relevant teams or stakeholders. This was a not so great thing. Once I discovered this, I set up several meetings with the involved groups to discuss everyone’s goals and expectations for the dashboard. These meetings surfaced knowledge gaps that were easily bridged with a meeting or two. After ensuring that each team was on the same page with regards to features, expectations, and design, we were able to align our goals and priorities. At this point, I was able to confidently synthesize the information so that I could begin further iterations on my dashboard designs. With a better understanding of the users' needs, the stakeholders' goals, and the future of this project, I was able to design an architecture that was durable and scalable and also accomplished the highest priority goals and expectations.