Syncing Content
All work
E-Commerce · Fashion2025
KAPAD
Rebuilt the catalog-to-checkout funnel for a premium fashion brand, recovering 40% in conversions and tripling page-paint speed.
- Client
- KAPAD
- Timeline
- 12 weeks · 2025
- Role
- Lead UX + design system
- Toolbox
- FigmaHotjarGA4Tailwind
[ 01 — Problem ]
Cart abandonment was 67%. Users dropped at the size-and-fit step, where the table was 6 columns wide and offered no fit guidance.
KAPAD is a premium clothing brand with rich photography and an ambitious editorial voice. The existing site looked great but converted poorly — heavy hero modules pushed product cards below the fold and the sizing experience created decision fatigue.
[ 02 — Research ]
How I learned what to build.
- User interviews · n=12
Recruited returning shoppers and lapsed buyers; recorded every pause longer than 3 s on mobile.
- Funnel analytics
Mapped 90 days of Hotjar + GA4 to isolate the size-step drop and product-image LCP delay.
- Heuristic + accessibility audit
WCAG AA pass; tap-targets, contrast, focus order. Fixed 18 contrast failures on the PLP.
[ 03 — Insights ]
- Insight 01
“Shoppers don't want more product info — they want fewer, clearer decisions. We cut PDP modules from 11 to 6.”
- Insight 02
“Fit anxiety beats price anxiety. Replacing the table with a 'find my fit' chat-style picker collapsed the highest drop-step.”
- Insight 03
“On mobile, hero imagery costs us above-the-fold conversion. We moved the catalog rail up by one full viewport.”
[ 04 — Process ]
Discover, define, design, ship.
- 01Discover
12 interviews, funnel teardown, competitive teardowns (Aritzia, COS, & Other Stories).
- 02Define
Reframed the brief from 'redesign PDP' to 'remove fit friction'.
- 03Design
Two PDP directions; A/B tested fit picker vs improved table on 18% of traffic.
- 04Ship
Componentized PDP modules in Figma, paired with engineering to map to Tailwind tokens.
[ 05 — Solution ]
The interface that shipped.
[ 06 — Impact ]
Numbers from production.
- 40%
- Conversion lift vs prior funnel
- 3×
- Faster paint 1.9 s → 0.6 s LCP
- −54%
- Size-step drop
- +18
- NPS post-purchase survey
[ 07 — Lessons ]
The biggest design wins on this project were subtractive. The Figma file ended the project smaller than it started. That ratio — components removed vs added — has since become a metric I track on every project.
Next case study
Read case studyIOTBIT
A calm operator dashboard for 150+ industrial IoT devices, dropping decision latency below 50 ms.