Alibaba Internship

Redesign retention of order cancellation for a food delivery app, which retains 97k+ more orders per day.
Mobile
Persuasive Design
Data-driven
DURATION
2 weeks, June 2021
TEAM
1 UX Designer (My Role)
1 Design Manager
1 PM
1 Illustrator
DS Team

Overview
Ele.me, one of the biggest food delivery platforms in China, had 25 million orders per day. However, there were 1.62 million (6.44%) orders were cancelled. One of the business goals of my team, the Order Transaction & Fulfillment team, was to reduce order cancellations. With this data-driven goal, I single-handedly delivered several design-initiated projects to improve fault tolerance, retain cancelled orders, and guide fast follow-up order. This case study focuses on retaining pre-cancellation orders -
How might we retain the order when the customer wants to cancel it?
By providing targeted retention designs and prompts that match users' motivation and ability to continue orders, my design launched in August 2021 has increased pre-cancelled orders retention by 6% and helped the platform retain 97k+ orders per day. Β πŸŽ‰
‍
‍
‍
My process
Apply design thinking in a business & data-driven setting
The Problem
Customers cancel orders for various reasons.
To dig into the problem behind order cancellations, I started with analyzing the existing cancellation reason data. Here are the two major insights I identified:
Since there were various reasons why the customer cancel an order, I had an initial thought that it might be better if the app had targeted strategies to deal with different cancellation reasons.
‍
‍
‍
However, the existing retention design was not persuasive and targeted.

CHALLENGEΒ #1
How might we provide more targeted retention design for different cancellation reasons?
Secondary Research
How could we change users' behavior?
To approach this problem, I did secondary research on how to change user behavior and found Fogg Behavior Model as a useful reference. This model states that prompts can change behavior only when both motivation and ability are high.
‍
‍
‍
‍
Competitive analysis
Order cancellation retention strategies in the industry
At the same time, I looked into the retention strategies of 11 apps across 3 related industries, including food & retail delivery, transport service, and e-commerce. I summarized the common strategies using the motivation-ability matrix, as a reference for further ideation.
‍

‍
‍
RETENTION STRATEGIES
Match reasons for cancellation with targeted retention strategies
‍
CHALLENGEΒ #2
How might we integrate the retention strategies into the current user flow?
SOLUTION #1
Put the reason selection page ahead
In the existing user flow, the retention message popped up before the reason selection page, so the system could not leverage the cancellation reason to provide targeted retention. By putting the reason selection page ahead, the popup would be able to show targeted retention messages and action entries based on the reason the user just select.
‍


‍
However, there were 2 critical concerns...
When I tested the user flow with 6 random users, all of them expected the order to be cancelled immediately after they selected a reason, so the retention popup was confusing and annoying to them. Thus, I rejected this solution, and came up with more ideas about leveraging the existing flow.



‍
SOLUTION #2
Provide different retention based on order status
‍

‍
‍
Validate assumptions with existing data
The idea was based on the assumption that cancellations due to users were more likely to occur in the first half of the journey, while cancellations due to merchants or delivery were more likely to occur in the second half of the journey.
‍

‍
‍
To verify the assumption, I turned to our DS team to find related data. Thought there was no detailed cancellation data related to each order status, the existing data showed that the orders cancelled before payment or within 1min after placing the order were mostly customer reasons, and cancellations due to merchant or delivery reasons were mostly occur after that, which supported my assumption to some extent.
‍

‍
‍
‍
SOLUTION #3
When selecting a reason, the corresponding guidance pops up below
‍

‍
‍
‍
Iteration
Solution #2 + #3 won by applying retention strategies while guiding users to take action!
Solution #2 would apply persuasive strategies in the popup, and solution #3 would guide users to take actions after they select a cancellation reason. So the combination of them would achieve 2 kinds of retentions with different focuses.
‍

‍
‍
However, it also had the problem of relying on user-selected reasons.
If the user select a random reason, then the app could not show the action entry that fits the user' real needs. To solve this problem, I was thinking about showing the action buttons such as "Edit Order Information" in the confirmation popup as well.
‍

‍
‍
Alert v.s. Bottom Popup
During this iteration, I also changed the UI from Alert to Bottom Popup, because the latter could provide more room for illustration and make the buttons more accessible. Also, the slide up interaction would be less disruptive than Alert.
‍
‍
Adapt to FeaSIBILITY
Prioritize the most frequent scenarios
In communicating with the PM and engineers, we prioritized the most frequent scenarios in the first sprint based on the current data.
‍
‍
‍
‍
Collaboration
Communicate with the PM, design manager, engineers, and illustrator to drive the design-initiated project and facilitate implementation
FINALΒ DESIGN
A set of targeted retention designs that persuade and guide users to continue orders
IMPACT
Achieved business goalsΒ  πŸŽ‰πŸŽ‰πŸŽ‰
‍
reflection
Learnings
  • Collaboration & proactivity in a large company
  • Leverage existing data
  • Ground design on solid theory and best practices
  • Consider realistic constraints - minimize cost & maximize impact
  • Problem solving in a complex system with multiple stakeholders
‍
What I would do differently?
  • Validate long-term effect
  • Consider possible different personas
  • Go deep in emotional design
‍