Contact Me

Use the form on the right to contact Amy Huang. Or email to amyhuangdesign@gmail.com

28 樂古道
香港島
Hong Kong

708-469-8117

Changing Seasoned Customer Behavior

WHAT DOES SHOPBACK DO?

ShopBack is an e-commerce platform that offers cashback and discounts, aiming to help users save money on various online shopping and travel bookings. It has over 30 million users across 13 countries.


SHOPBACK BUSINESS MODEL

BUSINESS PROBLEM

Shopee aims to tighten the affiliate criteria to reduce commission expenses in the affiliate market. Previously, users would simply redirect from ShopBack to the Shopee homepage. Now, they must identify the specific stores they wish to buy from on Shopee, search for these stores on the Shopee AMS page in ShopBack, redirect, and complete their purchase through ShopBack. This change poses a risk of losing millions in commission revenue if users continue their previous behavior.


OBJECTIVES & APPROACH

  • Identify challenges in the user’s journey when learning to earn Cashback through AMS (Affiliate Marketplace Solution) page.

  • Evaluate the effectiveness of communication materials in raising users' awareness about the changes.

RESEARCH METHODS


KEY FINDINGS

  1. Inefficient Banners

    Banners looks promotional are more ineffective, in usability testing sessions, users did not fully understand the message being conveyed under the banner even they stop and read it.

“Yes, I have seen, up to 10 percent more.... maybe if I buy this item [under Deals for you], I will get 10 times more Cashback.” - User 1

2. Broken Feedback Loop

Inconsistent cashback displays prevent users from feeling a sense of earning cashback, leading to a lack of repeated behaviour.

3. Minimal Cashback Expectations

Users anticipate minimal cashback due to longstanding expectations. Years of experience makes it difficult to shift perceptions and capture user attention to introduce new mechanic.

“I want to maximise the cashback, but I didn't get it, it is still okay. cuz the amount is small.” (User 3)

“I will get around 1 peso, like 0.1~0.2. For clothes, nike adida I don't buy from Shopee, I will wait for higher Cashback from ShopBack stores.”  (Pilot user)


Learning & Observation

AB testing process

Due to my previous lack of experience with A/B testing, this research helped me understand that users are automatically assigned to different versions of the A/B test the moment they open the app. Therefore, when conducting qualitative user research during an A/B test, if we start recruitment after launching the A/B test, we cannot pre-select users based on the version they see (e.g., recruiting users of version A or version B). Conversely, if we want a complete list of users for recruitment, we must wait until the experiment is finished. Given the time constraints of this project, we had to over-recruit users and only determine their assigned version when they arrived for the interview.

Minor Errors in Content Can Cause Significant UX Problems

During the research, we discovered that inconsistent information display often leads to larger systemic UX issues. Over time, users have come to distrust the displayed cashback rates, believing they are inaccurate. This seemingly minor issue results in a significant breakdown in the feedback loop, preventing users from feeling that they have successfully received their cashback.

Although we found that the display of cashback rates involves complex backend processes and requires cross-departmental collaboration, this problem has demonstrated how even small errors can profoundly impact user behavior. Users' lack of trust means they cannot close the feedback loop effectively, as they do not feel confident that they have received the cashback they were promised.