Musinsa

At MUSINSA, I worked across reviews, style communities, customer experience, discovery, and advertising products inside Korea's largest fashion commerce platform.

On paper, those looked like completely different domains. In practice, they were all connected by the same questions.

How do people discover things? What makes them trust something? How do taste, identity, community, and commerce influence decision-making?

I started on the Reviews team, working across PDP, PLP, and review creation experiences.

Fashion reviews behaved differently from typical commerce reviews. Users were not just checking product quality. They were looking for signals about fit, styling, taste, identity, and social proof.

This became more obvious as MUSINSA expanded into beauty. I helped design and launch beauty-specific review experiences, including category-specific satisfaction attributes, review guide restructuring, and migration of existing review data.

One of the more interesting problems appeared during MUSINSA's large-scale seasonal promotions. Promotional products were created with separate product IDs, which fragmented reviews across duplicate listings. Users hesitated, then left the app entirely to search YouTube or Instagram for external validation.

User testing and behavioural data made the pattern obvious. The absence of trust signals was directly interrupting purchase flow.

I worked on a system that connected related products and shared review data across them, allowing users to access existing trust signals without leaving the platform.

Later, I joined Snap, MUSINSA's style content platform.

The core problem was not a lack of content. Style content, creators, brands, and commerce all existed inside the same ecosystem, but they were not reinforcing each other.

Snap was an ecosystem where different groups with different motivations interacted constantly: users looking for inspiration, creators seeking recognition, brands wanting affinity and conversion, and the platform optimising retention and GMV.

Instead of approaching this as a content problem, I approached it as a systems problem. How do you design a structure where those participants naturally strengthen each other's experience?

That led to several initiatives: consolidating fragmented style content, improving recommendation logic using shopping context, building creator reward systems, designing personalized retention flows, and expanding short-form video experiences.

One personalised retention initiative improved weekly retention by 8 percentage points.

Users who engaged with style content showed 26% higher GMV than users who did not.

Later, I joined a cross-functional CX task force focused on improving the overall shopping journey.

As MUSINSA scaled, different product surfaces had evolved somewhat independently. Each team optimised its own surface well, but the overall experience was becoming fragmented.

Users were not experiencing the app as separate features. They were experiencing one continuous journey.

The task force focused on reducing friction between existing systems rather than building entirely new features.

I aligned directly with the founder twice a week to review priorities, discuss tradeoffs, and iterate on cross-platform improvements.

This was probably the first time I started thinking less about individual features and more about orchestration at the ecosystem level.

Most recently, I joined the Discovery team to work on premium advertising experiences for fashion and luxury brands.

MUSINSA already had strong performance-oriented promotional surfaces. What felt missing was a space for brand storytelling.

Luxury and premium brands were producing high-quality campaign content designed to communicate mood, identity, and cultural positioning rather than simply displaying products.

Because I had previously worked on style content, creator ecosystems, and short-form video products, I joined the Discovery team to help design more immersive advertising experiences together with the advertising business organisation.