Lyft is one of the fastest growing companies in the new “sharing economy.” The current rating system tacitly encourages users to rank their driver using only very good or very bad rankings, due to the known limitations of a five-star scale. As given in this class project design situation, our goal is to create a replacement for the current five-star rating system that Lyft uses to track and reward drivers in their system.
Time | 09,2017 3 weeks
My Role | User research, user interface design, usability testing
Lyft Rating System Redesign
The goal of our secondary research was to understand the current system from each stakeholder and figure out the task flow of the whole process.
The driver's information page will show up while waiting
The user book the cab
After getting out of the car, users can rate. The driver will get a feedback report (the picture on the left) if the user rated. The lowest score for the driver to keep the job is 4.8
After rating, users will pay the tips and submit payment
We chose the survey and semi-structured interview as the data collection method to get our primary data from users and created the affinity diagram to sort and analyze the data. Finally, based on the findings from both the secondary and primary research, we created two personas to solid in our minds the needs of the users.
The survey aimed to figure out the general views from users about the rating system and finally we got 35/55 responses, among which 32% used Lyft weekly. The key findings are as follows:
Users have the different understanding of the 5-star rating.
The goal of conducting the interview was to get the specific comments on the rating system. We interviewed 5 people including 1 driver and 4 passenger. The answers were audio recorded and transcribed. Also, some salient quotes were captured.
We transcribed all the data and used post-it-note to sort the data.
After sorting all the notes, we finally classified the information into 5 groups.
People's attitude about the ratings,
The preference people have in a car ride,
How people understand the current rating system
People's experience with Lyft
Primary concerns during a ride.
Two personas were created after the data analysis. Our primary persona was the user's side, while the secondary persona was from the Lyft manager side to help us better understand the user and the whole system.
Julia works in Go-service, a start-up which involves a lot of travel to new places for client meetings. Julia attends presentations in odd times and uses LYFT to shuttle between hotel and Presentations in a tired state. She mostly had good rides in LYFT and eventually got bored of the rating system. Now she only wants to point out if she has any problem with her ride and not otherwise. Sometimes, she finds the driver is not at fault for her ride to be bad and cannot really understand how she can modify that.
"What's the difference between 2 and 3? 4.8 and 4.9? I am so confused."
1.Julia wants to quickly rate
2. Julia doesn’t want to decide so much on which “Star” to give the driver
3. Julia wants to know the meaning behind the score
Sam is a developer in LYFT for more than two years. Sam is a system level manager running a team of 10 other data analysts. His team is in charge of the rating analysis. Recently, the team hired more analysts to form a new platform for semantic analysis of rating to improve the performance of the LYFT System. Sam is in contact with the design consultancy for trying to get appropriate user inputs.
' I can hardly get the data of comments on low ratings."
1.Sam wants right data points for data analytics
2. Sam wants to build a system to qualitatively analyze the rating that user gives
3. Sam wants to submit reports of the analysis to LYFT to improve the overall system from the results of the analytics
26yrs| Female| Product Manager
26 yrs| Male|Data Analyst
The current rating system cannot reflect the user's real problem in a car ride experience, at the same time, users are lack of knowledge of how to rate. In such case, drivers cannot get appropriate feedback on how to improve and even having the chance losing the job due to user's lack of knowledge of the rating system.
The Design Goal
Users know how the rating system works and how it can benefit them.
Users can express the problems easily and precisely.
Using emoji and color to indicate the meaning of the score
slider bar to emphasize the sense of manipulating
Adding visual representations to show what constitutes the final score on driver's info page
The rating page.
The driver information page
Our solution can be divided into two parts. The rating page and the driver's information page.
For the rating page,
The big idea behind the four slider bars was to enable the user only to mark the categories they would like to highlight so that they can express the problem more precisely.
we integrated the elements of emoji and color indicator in the slider bar. Through smile and sad face as well as the change of the color people can get the sense of whether the score was too low or too high.
The driver information page
This was the first-page user would see when the cab was booked, instead of just showing the rating score, we changed the user's mental model by showing the components of the score first, thus users can not only know they can benefit from rating but they would know the meaning better.
Our second iteration was based on the findings of the cognitive walkthrough.
For the rating page,
We changed the emoji icons to the current ones which were corresponding to the four categories provided by Lyft.
The starting position of the icon was moved to the middle to avoid misunderstanding.
Also, we changed the bar by adding a sense of being selected.
The wording was changed
In the driver information page
we changed the visual representation and the icons used for facilitating the understanding were added.
The final edition was done after 2 rounds usability testing. We tested with 5 users in total.
Driver's information page
Instead of just showing a rating score, we provided a visual representation of the components for the final score.
Users will get the first impression of how the rating works and get some useful information from it.
From our results of the usability testing, we got to know that some users would like to see some detailed information so we created a subpage for more info.
For the users: Have a better understanding of the rating and have more information to decide whether take this cab.
For drivers: Good understanding of the score of users will ease the driver's burden to educate the users.
For the system: To help monitor which cab is being frequently canceled.
The completely new rating page
We changed the current 5 - star rating system to this new one.
The user can mark the categories that they want to highlight by clicking to activate the bar.
Scrolling up and down the emoji icon to rate, the emoji and color of the icon as well as the text below the score can help users understand the score better, thus they can give the rating precisely and easily.
If the user did not rate, the score would keep the same as the driver's current score ( based on the last 100 rides).
For the users: Just need to give rating on specific aspect
For drivers: The specific score can definitely benefit the driver by letting the driver know which aspects to improve easily.
For the system: Monitor and analyze what a driver has to be specifically trained about.
The separate comment page
From the research, we got to know that the comments from the users were quite important to drivers because this was the key source to avoid biased ratings and to know the details of how to improve. Sometimes the low rating was not caused by the driver but the traffic or other factors.
To stress the comments we created a separate for it can some comments from other users were provided to motivate the users to rate.
For drivers: Get useful details to improve and avoid bias
For the system: To monitor the drivers using more precise data.
The design is never finished, never complete. Every design project is an iteration process.
In this project, we had 3 iterations in total and used usability testing, heuristic evaluation as well cognitive walkthrough to inspect problems.
Although our project was finished, we were still eager to know whether this design can have some good or bad influence on passengers and drivers. We presented our design to the passengers and driver for feedback and happy to see that 2 passengers told us that they had a better understanding of the rating system through the new design, also,
from their comments, we get some insights on future work:
From the driver's perspective, they do hope they can also know more about the passengers. The comments left by the drivers cannot reflect what the passenger looks like.
The time spent on the whole process needs to be faster