Customer Discovery Platform

The Challenge

With millions of SKUs, customers struggled to find what they needed. Search was strong, but browse and discovery were weak. Category pages felt like faceless grids with no editorial voice or personalization.

My Role

I led a team of four designers and one researcher. I drove the product vision, ran design reviews, and partnered with ML and merchandising to align on personalization strategy.

Approach

We conducted extensive user research to map discovery behaviors. We designed a flexible card-based layout that could support editorial, algorithmic, and hybrid sections. We A/B tested different recommendation strategies and iterated based on engagement and conversion signals.

Outcome

• Engagement with discovery surfaces increased 42%
• Time to first product view decreased 3x
• Customer satisfaction for “finding what I need” improved from 67% to 81%

Learnings

Discovery is not search. It requires a different mental model: curation, serendipity, and trust. Balancing algorithmic relevance with editorial voice was the key to making it feel human.