ZALORA / Experiment

Elevate Cart Experience Through Experiments

MY ROLE
Design
Research
PLATFORM
IOS & Android App
PROJECT DURATION
4 weeks
YEAR
2023
PROJECT summary
ZALORA's conversion rate dropped, and we urgently need to develop high-impact strategies to boost it by at least 0.25% and get it back on track.

I drove lean research to rapidly iterate through A/B testing to optimize cart to order completion flow, resulting in 1.5% increase in conversion rate and 15% boost in average order value.
BACKGROUND
Last year, ZALORA revamped the cart and checkout design, which performed well in A/B testing at the time. However, by mid-year, data showed a drop in our Conversion Rate. Based on data analysis, our team identified that even a small improvement in the cart-to-checkout Click-Through Rate could positively impact the overall CR. To address this, we urgently need to come up with high-impact strategies to boost the conversion rate and get it back on track.
BUSINESS GOAL
Increase the cart-to-checkout CTR to improve the CR by at least 0.25%.
Increase Average Basket Size
Reduce Cart Abandonment Rate
current DESIGN
The checkout process consists of 3 steps: cart, payment and review. To streamline the work, my focus in this project is the cart and checkout page.
APPROACH
In collaboration with the PM, we took a data-driven approach by breaking the issue down into key areas for analysis. From there, we used an experiment-based strategy to develop hypotheses, identify improvements and benchmark against competitors in the market.
DESIGN DIRECTION
🪄 Assisted by having smart actions and suggestions to minimize manual work for users.
🕹️ Adequate by bring understanding without overwhelming users, facilitating informed decision-making.
HYPOTHESES
Design Decision:
I changed the prompt colour from red to green to make it feel less like an error, while keeping it noticeable and aligning with the delivery information colour.

On the recommendation page, I wanted to reduce the effort for users to find relevant products. Since adding filter and sort options would extend development time, I decided to arrange wishlisted items at the top, sorted by how close they are to meeting the minimum spend requirement.
Design Decision:
This message will only appear when X users have added the item to their bag or wishlist, showing that there's interest from others. However, if the item is low in stock (2 left, indicated by a red label), this message will not be displayed.

For placement, I positioned it right after the product name to ensure it doesn’t disrupt other related details.
Design Decision:
With the goal of highlighting product discounts, I placed this information under the voucher section and changed the title from `Total Discount` to `Total Saving` to convey more positive message.

Ideally, I’d like to display the total of the product discount and voucher together, so the savings amount appears larger, but due to development effort, I’m going with this approach instead.
Design Decision:
I decided to test both the floating bubble and sticky bar because these two components provide users with quick access at any time. The floating bubble is positioned on the left side for better thumb-friendliness.
Design Decision:
I added an animated icon and made the saving amount really pop to grab users' attention and highlight their savings.

As the data shows that users frequently navigate back and forth on this page, I want to ensure they aren't bothered. So, I kept the modal to appear only twice per session and added a skip option. This allows us to monitor any potential annoyance for users.
USABILITY INSIGHTS
After quickly coming up with the design, we conducted a rapid concept test with our power panel users, which uncovered valuable insights. We used these insights to further refine the design before finalizing the experiments.
FINAL DESIGN
We breaking down each hypothesis and A/B test them with different groups to determine which one achieves the desired impact.
RESULT
+1.5%
Conversion rate
+15%
Average Order Value
Above are the total impacts from the experiments and below is the individual impact for each hypothesis.
BACK HOME