How does keeping versus removing a fixed discount rate of 25% affect overall sales performance?

The complexity of the price system was mainly attributed to a "fixed shop discount of 25% off," a promotional method that was separate from the coupon causing users confusion of price logic. To weigh the effects of keeping vs removing the discount on sales performance, we launched a 25-day A/B test for all users.

As the new version without the fixed 25% discount generally outperformed the original version, we decided to remove the fixed discount, presenting users with a simpler, more transparent price system.

Building components with only necessary variants to tackle the inconsistency in content, visual style, and structure

We tested the redesign prototype on 5 participants, achieving an SUS score of 82.5: a significant increase from the previous score 59.17.

Take all edge cases into consideration and build an organized system

In this project, I realized the heavy inconsistency problem resulted from absence of a clear system that structurally listed every edge case--in some situations, it took me much longer to figure out how the existing system works than designing a new one. Real-world problems can be very complex. In e-commerce settings, all the combinations of purchase options, order status and market strategies would be a great challenge, and our job is to make sure our design has taken every case into consideration.

Make fully-informed design decisions

Design decisions cannot be based on the UX designer's "intuition" on what would be good for the users. Our hypotheses should be supported by solid evidence, which includes drawing insightful conclusion from in-depth analysis of data, user's feedback, etc. As a D2C company, our strength is having direct access with users' data and feedbacks, which allows us to make sure we are spending our limited resources on the right thing.

Communicate with stakeholders using data, business impact

We can't always agree with each other in the process, but we can get on the same page using recognized tools. When we were defining the scope of design, the RICE model were helpful in combining every one's opinion from different aspects (cost, business impact, confidence in success), and giving us a final score that the team would accept. The key in communication with stakeholders or teammates is similar: embrace broad perspectives, and decide by recognized evaluation.