MY ROLE: research alternate ratings systems, create sketches and focus on the "first impression rating screen" low-fidelity wireframes, test with users
GOAL: create a replacement for the current five-star rating model that Lyft uses to track and reward drivers in their system
Rate Your Lyft
Fall 2017 - Purdue University
Lyft is one of the fastest growing companies in the “sharing economy.” One of the most important aspects of the sharing economy is qualified and accurate ratings from users, which allows operators to divert resources appropriately and reward contractors that provide high levels of customer service. 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.
PRIMARY + SECONDARY
USER TESTING +
REVISED DESIGNS +
PRIMARY + SECONDARY RESEARCH
We recruited five participants for semi-structured interviews and a design workshop.
We recruited five participants for semi-structured interviews and a design workshop. Some notable quotes from interviewees are listed below.
"Ratings above 4.5 are all the same to me, there are always people who will rate badly for a bad reason that they found irritating, but isn’t pertinent to the safety or destination."
"I don’t check the driver’s rating. Lyft is supposed to deal with the issue,
or it will harm the company’s reputation."
We researched the current Lyft system, including the inner workings of the company, as well as alternative ratings system designs.
"Anything more than 4.8 is awesome. If your rating drops below 4.8, you may want to consider ways to improve it. Consistently low ratings can put you at risk of deactivation." - Lyft Help Center
"Users’ rating costs increase as they have more rating choices. All scales show similar cognitive load. However, page rating times increase significantly with finer-grained scales - despite users leaving more items unrated with those scales." - Rating: How Difficult is It? (Sparling & Sen)
After gaining insights from primary and secondary research, we arranged our findings into common themes.
Our affinity diagramming was iterative; we moved from smaller sections to larger, more inclusive categories. We also organized information chronologically, designating categories under "Before," "During," and "After" the Lyft ride experience. Important insights from affinity diagramming include the importance of the rider's safety, the desire for a simple rating system.
We created two personas based off of our primary research. Each persona has a differing personality regarding their preferences during a Lyft ride.
“I care about safety and I trust Lyft since it inspects the drivers and tracks my route. The best thing is that the drivers are always kind and supportive.”
When traveling, if Molly needs to move from place to place, the first thing comes to her mind is Lyft. She chooses the basic option instead of Plus, Premier, and Lux since she usually travels alone or with a small group and this option is always the cheapest one. During the ride, Molly enjoys seating in the front passenger seat and talks nonstop with the driver. As long as the driver takes her to the destination safety within a reasonable time, she doesn’t really notice driving behavior. Molly always asks the driver for anything special in the city and casually shares each other’s life experience. She thinks being able to talk with the locals during the rides is a benefit of taking Lyft.
Outgoing, Sociable, Communicative
Molly Hunter, 22-year-old student
- quick rating system
- chatting with driver
- clarification about what ratings actually
- a space for comments
- confusion surrounding what a five star
ride is, what a one star ride is
- doesn't always have time to rate
“I would normally just sit quietly in the back seat and relax, thinking about things I have to do at my destination.”
Soo-Jung Park, 26-year-old graduate student
- quick rating system
- fast and direct service
- a fast way to leave feedback on his ride
to prepare others for this driver
- uncomfortable with English, and writing
- worried about safety
Occasionally, he would use Lyft to get around when he travels to larger cities where car rental costs and parking fees are somewhat high. He is a bit shy and does not speak perfect English. Thus, when he takes a Lyft, he would normally sit in the back and prefer not to initiate a conversation with the driver unless he has questions to ask the driver. Normally, he would leave a 5-star rating to most drivers. If he is not completely satisfied with his trip, he would just force quit the app, and do not rate the driver.
Introverted, Hardworking, Quiet
We created sketches as a group, and also individually, coming together to compare and vote on ideas.
Initial ideas included an emoticon rating system, a three-tag rating system (good, neutral, bad), and a binary thumbs-up, thumbs-down system, and an inclusion of a quick swiping motion similar to a Tinder rating. Secondary research created strong support for a binary system. Our ideas were diverse, so we tested three different binary system designs with users.
USER TESTING + FEEDBACK
We tested 3 different rating wireframes with 4 participants asking them to think aloud, testing task flow and usability, as well as different riding experience scenarios. We also presented the participants with variations of a driver profile wireframe, asking participants to share their thoughts and preferences.
In general, there was consensus among participants that Wireframe B was the
preferred option. There was resistance from participants when presented with multiple consecutive screens to give ratings, and there was confusion surrounding the placement and aesthetics of the thumbs-up and thumbs-down buttons. Participants liked the overall binary rating system for its simplicity and efficiency, but did not enjoy rating individual aspects of the design, saying it was too much effort.
REVISED DESIGNS + PROTOTYPING
In order to address the problems revealed through user testing, we continued to iterate upon our designs to create a final design. This design combined some elements from our initial wireframes, modified some, and discarded others. This final ratings design addresses our problem statements of an inefficient and ambiguous rating system, as well as a limited amount of feedback to the driver.
The user is presented with a binary rating system to rate the overall ride first. After rating the overall experience, tags appear, and when selected, signify the good aspects of the ride. Comments are optional, and users can opt out of rating any of the experience altogether.
The user is presented with a binary rating system to rate the overall ride first. After rating the overall experience, tags appear, and when selected, signify the negative aspects of the ride. Comments are optional, and users can opt out of rating any of the experience altogether.