The Challenges

1. Traffic was scaling faster than the site could convert it: Paid campaigns kept sessions growing, but revenue growth depended on the store converting that traffic better, not just buying more of it.

2. Weak product discovery on a mobile-heavy store: Around 9 in 10 sessions came from mobile, yet desktop converted roughly twice as well. Shoppers looked for specific product types (kurung, kebaya, kebarung) but navigation and search made them dig for it.

3. Extreme festive seasonality: Close to half the year's online revenue arrives around Raya. Every conversion improvement had to be validated and live before the peak window, or wait a full year for the next one.

4. A low returning customer rate: Repeat purchase behaviour lagged behind the brand's offline strength, leaving revenue dependent on constantly acquiring new buyers.

‍

Key Objectives

1. Improve Product Discovery & Navigation

2. Lift Add-To-Cart & Checkout Completion

3. Grow Average Order Value (AOV)

4. Turn First-Time Buyers Into Returning Customers

‍

The Solution

1. No big-bang redesign: 22 months of continuous CRO, averaging 9 to 10 research-backed improvements shipped per month. 90 optimisations live, 15 A/B tests concluded.

2. Navigation and search rebuilt around how shoppers actually browse.

3. Cart and checkout flow re-architected step by step for mobile.

4. Every meaningful change A/B tested. Winners scaled, losers rolled back.

‍

Use Case 1: Predictive Search That Matches How Shoppers Browse

Search was one of the highest-intent behaviours on the site: shoppers who searched added to cart 63% more and reached checkout 43% more than those who didn't. But tapping the search bar gave them a blank box. We introduced predictive search terms so popular searches surface the moment the search bar is opened, guiding shoppers straight to what most people are looking for.

Results:
+17% search usage
+10% AOV from search sessions
+26% incremental revenue through search

‍

Use Case 2: Replacing The Cart Drawer With A Full Cart Page

On a store where nearly all revenue comes through mobile, the slide-out cart drawer compressed the free-shipping progress bar, order notes and cross-sells into a cramped panel. We tested sending shoppers to a dedicated cart page instead.

Results:
+10% checkout rate
+4% AOV
When we retested the same question two years later, the cart page won again.

‍

Use Case 3: The Test That Failed, And Why It Mattered

Behavioural data showed the product page's add-ons section earned far fewer clicks than the recommendations section below it, so we tested removing it to shorten the page. Add-to-cart dropped 15%. The section was quietly doing more work than click counts suggested. We rolled the change back within the cycle and optimised the section instead of deleting it.

On a store where close to half the year's revenue lands in one festive window, shipping an unvalidated change like this before Raya would have been expensive. Testing is what caught it.

‍

The Outcome

Comparing the Raya quarter year on year: sales grew 2.1x while traffic grew 1.8x. The gap between those two numbers is where the conversion work shows up.

+82% orders
+18% AOV
+12% checkout completion rate
+24% returning customer rate
+18% returning customers' share of sales

Paid traffic scaled in the same window, so no single lever takes full credit. What the programme changed is how much revenue each visit produced, and who those visitors were: by the end of the engagement, a meaningfully larger share of sales came from customers who chose to come back.