Bridging Stakeholder Interests
March 15, 2026Tasks
Research, UX / UI, Implementation
Tools
Figma, After Effects, Qualtrics, Maze.co
The Mission
Design and validate a new recommendation feature for an international e-commerce giant in South America and South Asia.
The Core Challenge
Increase "items per order" by introducing a new Market Basket Analysis algorithm without adding friction to the "Buy Now" journey.
Key Outcome
60% higher conversion preference for the new algorithm and positive user sentiment toward a new "Interstitial Recommendation" pop-up.
Context & Constraints
- The Problem: High shipping costs often exceed the value of small (1-2 item) baskets, leading to cart abandonment or low profitability.
-
Technical Constraint: Must handle "Multiple Seller" shipping complexities and product variations (size/color).
-
Target Audience: High-speed shoppers who often skip the cart page entirely.
Strategic Discovery
Competitive & Journey Mapping
-
Competitive Insight: Top players vary recommendation visuals based on where the user is in the journey. This informed our decision to test different UI for PDP vs. Cart.
-
The "Gap" in the Journey: Mapping revealed that users skipping the cart missed all current recommendations. This led to the "Buy Now" Pop-up hypothesis.
The Solution: A Two-Pronged Approach
-
Integration A: The Passive Approach (PDP & Cart)
-
Implementing the "Statistical Affinity" algorithm in standard feeds.
-
-
Integration B: The Active Approach (The Interstitial Pop-up)
-
The Bold Move: Adding a recommendation step after clicking "Buy Now" but before payment.
-
Design Fix: Addressed the "Variation" problem by allowing users to select size/color directly within the recommendation widget to prevent back-and-forth navigation.
-
Validation
Unmoderated Test on Maze.co
"Data over Opinions" — We tested 19 users to validate if the new algorithm actually performed better than the legacy one.
| Mission | Goal | Result |
|---|---|---|
| Algorithm Head-to-Head | New vs. Old Algorithm | 60% preferred the New Algorithm's suggestions. |
| Placement vs. Product | Does page position matter? | No. Users chose based on relevance, not how much scroll was needed. |
| The Friction Test | Did the Pop-up annoy users? | Positive impact. Users found it helpful, not intrusive. |
Final Impact & Reflection
-
The Win: Proved that a "disruptive" pop-up can actually improve UX if the recommendations are highly relevant.
-
Business Value: The new algorithm is set to increase "items per basket," directly tackling the shipping-cost-to-item-value ratio.
-
Personal Growth: Delivering a full research-to-prototype cycle in under 14 days required high-tempo collaboration with the BI (Business Intelligence) team.
