Project Context

Recently, there is a constant voice online asking Uber to develop a feature that allows riders to favorite drivers or request specific driver. My capstone team would like to dive deeper into this user request and figure out the actual "user needs" underneath the "user wants".

Executing independently, we conducted user research and found out that users actually need more freedom to customize or recreate a memorable Uber experience, rather than to request a specific driver. It's more about the entire experience, the car, the service, the conversation, less about the driver personally.

Therefore, our design question was that "How might we help Uber riders recreate an Uber experience?"

TeammatesHyewon Son (Designer), Rachael Cicero (PM), Christina Mauri (Researcher)
Sponsored by: Uber

Methods & Tool
User Interview, Competitive Analysis, Usability Studies, Affinity Diagram, Figma

Role
‍Product Designer

Date
October 2018
- March 2019

Process

Followed HCD process throughout 6 months

We followed the User Centered Design process to investigate how Uber riders may want to recreate Uber experiences based on their preferences. Letting user research be our guide, we brainstormed and designed multiple solutions. We then tested our ideas via usability tests and developed a high-fidelity interactive prototype based on our findings.

Discovery

Competitive Analysis: similar products already existed

To start with the research, we analyzed 4 direct competitors in the ride-sharing industry (Lyft, Wingz, Didi, Uzurv) and 1 mobile app that serves two-sided market (Wag!).
We learned that:
1) Wingz, Uzurv, Wag! allow users to schedule with preferred drivers/walkers. Users can choose specific drivers as well.
2)The list of preferred drivers, in most of the applications, are the ones users already had before. The users can either mark them as favorite or add by inputting driver id number.
3)Uzurv gives riders to search for drivers with specific amenities and favorite them. The driver profile on Uzurv is richer than that of Uber.

Learned from Uber employees: scheduled ride would be a better use case

We also conducted 3 subject matter expert (SME) interviews with Uber employees to learn about Uber’s business and product goals. These interviews helped us determine scope for the project and to better understand the opportunity for a favorite driver solution. We learned that matching with specific drivers or types of drivers will lead to longer waiting time and higher trip price. Whereas in the Scheduled Ride scenario, riders will wait up to 30 days for the upcoming trips, so it makes more sense for system spending more time matching specific drivers based on riders' preferences in that case.

Visiting Uber Greenlight Hub in Seattle

User Interviews: wants ≠ needs

According to Uber employees' suggestions, we interviewed 8 Uber Premium riders to collect qualitative information about the rider and driver experiences and to gauge whether riders and drivers had a need that might be met with a favorite driver feature.

Conducting user interviews and affinity diagraming

Use Cases

Identifying high-impact scenarios when customization is needed

While analyzing our findings from user interviews, we identified airport rides and special occasions to be the most common use cases or scenarios where riders stated they would utilize a solution to customize their rides. Therefore, we determined the solution needs to account for:
1. High-stress situations 2. Time sensitivity
3. Quality of service 4. Safety

Design Requirement

People want "the least number of taps"

Thinking through what a rider would need during the use cases identified, we determined design requirements for the solution. We did so to ensure we were holding ourselves accountable to referring back to the data learned in research, and to use these requirements as our guiding principles as we moved onto brainstorming.

Ideation

Brainstormed the possible UX for the new feature

Designing a new feature within the Uber rider experience required us to brainstorm not only the design but how we would introduce this new flow to users. We explored different solutions and sketched out the user flows. Then we discussed and decided the most feasible solution for this new feature.

Solution

Favorite Driver Solution

We identified two possible design solutions. The first, an option for riders to input certain preferences (i.e. a quiet ride, extra luggage, room for pets) for each ride. The other, an option to favorite and request specific drivers that delivered an ideal Uber experience. We chose the latter, as we thought by allowing riders to request specific drivers, they would really be favoriting an entire experience which embodied their preferences, and giving them an opportunity to recreate this experience.
‍
We named the solution Favorite Driver.

Service Design Blueprint: How does it work

To help us identify current gaps in the ride experience, we created a service design blueprint. Through this exercise, we identified new entry points for the favorite driver solution. In addition, we were also able to identify both front end and back end processes needed to support this new feature.
We then created a second service design blueprint to help us identify current gaps in the scheduled ride experience. Coming out of research we heard riders speak to wanting to use this feature in a scheduled context. We identified new entry points for the favorite driver solution, as well as a new workflow between the rider, driver, and backend system.

Usability Tests

Evaluate the mid-fidelity prototypes

We narrowed down on potential ideas and started prototyping Favorite Driver based on the flow identified in the service design blueprints. We created a mid-fidelity prototype and moved on to evaluate on that.

Using paper prototype for usability studies

We recruited a total of 6 Uber riders from our social channels. Through use of a screener, we screened for riders who take rides at least 3 times a month and had a mix of experience in scheduling rides in advance. Sessions were conducted at the University of Washington. Each session lasted around one hour.
‍
Our goals for the testing were to:
1. Test discoverability of newly designed components such as “Add Favorite Driver”
2. Test discoverability and usability of new interaction flow
3. Learn about riders’ behavior, especially around how and when they rate drivers

Iterations

Based on the key findings we synthesized through conducting affinity diagram, we iterated on the design and upgraded the prototype from mid-fidelity to high-fidelity.

Synthesizing insights from usability studies

Key Finding #1: Riders don't always give a complimnet even when they liked the driver

Originally we thought 5 stars, a compliment and a tip should activate Favorite Driver feature, but some users don't give a compliment even they liked the driver. This made us reconsider the triggers for the Favorite Driver feature. We decided to remove compliment from the requirement so that riders don’t miss any opportunity to use the feature.
‍

Solution - Remove Compliment from Favorite Driver Prompts

Key Finding #2: Managing a list of favorite drivers is burdensome

In our first iteration, we built the ability to customize Favorite Driver lists based on manual creation of categories to filter drivers (much like a custom playlist). However, through evaluation, we found that users felt this required too much effort and they would not utilize a feature like list.
‍

Solution - Replace Custom List with Compliments

Instead, we utilized compliments (an existing feature) as a default mechanism to help users record traits/characteristics of the driver. Keeping in mind not all riders compliment upon rating a driver, we allowed for the ability to add custom compliments at a later time to help recall a favorited driver.

Key Finding #3: Riders want information that can help them remember why they marked drivers as their favorite

In the test, riders mentioned that they would like to see more personal related information of the specific driver. Things like ride history and past trips might help. In addition, riders also pointed that the three-dot icon was confusing and they had a hard time recognizing the badge icon for each driver. When prompted what information would be helpful for riders to recall a driver, one participant said "something like a hashtag would be super helpful".
‍

Solution - Add more relevant driver information

Therefore, we removed driver ratings, since riders only care about their personal experience with the driver, not an overall average rating. Then we added previous trips, total number of trips, and last trip to help riders recall their ride. We also made compliments more legible by using text instead of the badges that currently exist.

Key Finding #4: Riders were confused about selecting specific drivers during scheduling

All of the riders were confused with the copy “Select Favorite Drivers” and the avatar photos during evaluation. One rider stated, “editing is weird when I haven’t selected anything.”
‍

Solution - Change wording and remove visual hint

Therefore, we tried to make the section less prominent by removing avatars and replacing “edit” with an arrow, so that it’s clear that riders don’t have to take any action unless they desire to. Also, we added help text below to let riders know that all Favorite Drivers are selected by default.

Key Finding #5: Riders don’t always have time to rate the driver after a trip

Some riders mentioned that they don’t always have time to rate the driver directly after a trip. Therefore, we added an opportunity for riders to favorite driver during the trip, when they’re closer to the destination. We did not want riders to potentially miss out on the favorite driver feature purely due to lack of time. Through competitive analysis, we found that “Lyft riders open the app during 53% of trips”, and we wanted to use the during-trip screen as a opportunity for interaction with Favorite Driver.

Solution - Add During Ride Scenario

Final Design

Our final design encompassed four main opportunity areas for rider interaction with Favorite Driver, within the Uber application. We updated one existing area, the rating screen, adding in a trigger and prompt for Favorite Driver. For the other three areas we were able to innovate, creating new experiences through our solutions for during-trip ratings, favorite driver selection in a scheduled ride, and the Favorite Driver list. Finally, though out of scope our project, we did explore and design a solution for driver acceptance. Thinking about original design requirements, we created a mechanism for driver’s to opt-in to Favorite Driver.
‍

Add Favorite Driver & Schedule Future Rides

Add favorite driver during a trip

Add favorite driver after a trip

Request favorite drivers for future scheduled rides

Manage Favorite Driver List

Add compliments for future references

Check out ride frequency and past trips

Remove driver from the list

Driver can choose to accept or decline a scheduled ride

Driver Interaction

Though initiated by the rider, the favorite driver solution is a transaction between the rider and the driver. The system notifies drivers when a ride has been scheduled, requesting them as a the driver. Once the driver receives the notification (the next time they are online) they have the opportunity to opt-in, either accepting or denying requests. We believe this will give drivers more control and consistency over their Uber experience and impact to daily earnings by being able to better plan out their days in advance.

Poster & Video - Check them out!

Reflections

Some potential risks remain to be considered

It is possible that rider prejudice (on the basis of traits like gender, ethnicity, age, or sexual orientation) will be reflected in riders’ repeat driver requests. As such, drivers that are of a certain demographic might receive more business than other drivers.

In addition, by creating a mechanism for drivers to learn the habits and whereabouts of repeat riders, unscrupulous drivers might take advantage of riders. For example, by knowing that a rider travels to the airport on a regular basis, a driver would know when they are not home, leaving the rider’s home and belongings more vulnerable.

Another potential risk is that when a driver drives a rider repeatedly using the Favorite Driver feature, the two might be more likely to do business without using the Uber app. As such, the driver and rider would benefit from the relationship, but Uber would not profit from it.

Updates

The feature was shipped by Uber 🚀

It was so exciting to learn that the "Favorite Driver" feature has been shipped by Uber in 2020, with most of our designs and idea. Feel free to find more information in this article and Uber Help page.